Step-up after LCR
Re: Step-up after LCR
Hi WE300
In attach another test with only the LCR stage
The previous one ws with Zsource 600 ohm and Zload = 600 ohm, 10k. 47k and 100k
This one is with Zs = 20 ohm, 100, 200 and 60 ohm and Zload = 600 ohm, as Tango specs
In attach another test with only the LCR stage
The previous one ws with Zsource 600 ohm and Zload = 600 ohm, 10k. 47k and 100k
This one is with Zs = 20 ohm, 100, 200 and 60 ohm and Zload = 600 ohm, as Tango specs
- Attachments
-
- Tango freq_answer_Zs_600_Zload_600_Vin200mV_Vout_10mV.png (59.05 KiB) Viewed 127765 times
-
- Posts: 2091
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Re: Step-up after LCR
WE300B,
I think the important factor here is to either drive the LCR with its characteristic impedance OR load it with its characteristic impedance and not both. If you choose to drive it with its characteristic Z then the output can (should) be left unterminated and if you load it with its internal impedance then it should be driven from as low a source Z as possible.
With either the driving/terminating Z in place, altering the other has little effect on the absolute frequency behavior and only shows the gain loss found by decreasing the raton of R-source to R-Load.
In the case of your 600Ω reflected from the primary of a transformer, in a perfect world it would be a precise pure resistance but in reality there are chances of slight deviation at the frequency extremes so you have to try it and measure.
dave
I think the important factor here is to either drive the LCR with its characteristic impedance OR load it with its characteristic impedance and not both. If you choose to drive it with its characteristic Z then the output can (should) be left unterminated and if you load it with its internal impedance then it should be driven from as low a source Z as possible.
With either the driving/terminating Z in place, altering the other has little effect on the absolute frequency behavior and only shows the gain loss found by decreasing the raton of R-source to R-Load.
In the case of your 600Ω reflected from the primary of a transformer, in a perfect world it would be a precise pure resistance but in reality there are chances of slight deviation at the frequency extremes so you have to try it and measure.
dave
Get Your Fix
www.hifiheroin.com
www.hifiheroin.com
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
Hello Walge,walge wrote: ↑Mon Jun 28, 2021 9:54 am Hi WE300
In attach another test with only the LCR stage
The previous one ws with Zsource 600 ohm and Zload = 600 ohm, 10k. 47k and 100k
This one is with Zs = 20 ohm, 100, 200 and 60 ohm and Zload = 600 ohm, as Tango specs
lcr risp Zs_20_100_200_600_zload_600.jpgTango
These are real test and not simulation.
You can, in this case, see the difference about network only.
I decided to use this last config just because I can have different Zsource using different tubes on CF and the quality of the freq. answer is always fine.
To get a perfect frequency answer ( as diagram I sent previously) you can trim a little with a Zload
Other diagram is a Tango original EQ-2L freq. answer, with Zs=600 ohm, Zload= 600 ohm.
Vin =200 mV
Vout= 10 mV
The attenuation at 1 kHz ( as a reference for test) of a network with Zsource=600 ohm and Zload = 600 ohm is 26 dB
freq_answer_Zs_600_Zload_600_Vin200mV_Vout_10mV.png
Thanks very much for posting this test data. The results look very encouraging.
I intend to drive my 600 ohm LCR network with a low impedance source, using a Lundahl 1692A plate-to-line transformer. I am not too concerned about the absolute impedance of the source, as I will be terminating the LCR network in a 600 ohm load impedance, which is a Sowter 9063 transformer, terminated in an impedance that will reflect 600 ohms back to the network.
With Best Regards,
WE300B
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
Hi Dave,dave slagle wrote: ↑Tue Jun 29, 2021 11:33 am WE300B,
I think the important factor here is to either drive the LCR with its characteristic impedance OR load it with its characteristic impedance and not both. If you choose to drive it with its characteristic Z then the output can (should) be left unterminated and if you load it with its internal impedance then it should be driven from as low a source Z as possible.
With either the driving/terminating Z in place, altering the other has little effect on the absolute frequency behavior and only shows the gain loss found by decreasing the raton of R-source to R-Load.
In the case of your 600Ω reflected from the primary of a transformer, in a perfect world it would be a precise pure resistance but in reality there are chances of slight deviation at the frequency extremes so you have to try it and measure.
dave
Thanks very much for the reply.
Yes, I think that you and I are thinking along the same lines here.
I will be driving the LCR network from a relatively low-Z source, that being a Lundahl LL1692A plate-to-line transformer.
As I had mentioned earlier, the network will be terminated in the reflected Z provided by the termination/grid-leak resistor on the secondary of the Sowter 9063 transformer.
You make a good point in that the grid leak resistor terminating the secondary of the 9063 may have to selected via actual test, to obtain a 600 ohm reflected Z on the primary of the transformer, so the LCR network is looking at the optimum load Z for the most linear amplitude response.
Best Regards,
WE300B
Re: Step-up after LCR
to WE
First of all do you have a proper test set?
Only following the results of a test lab you can reach the right results.
If you check the pdf from Docali :
Practical Considerations
on Designing and Measering
LCR RIAA Filters
Dr. Bernd Ahlswede (May 2018)
it gives you the right direction, with a good sound card and a proper setting you will have a good start point.
In addition, check each stage one by one regarding freq. answer and distortion and gain.
The final results is strictly related ot the error on each stage, it is the sum.
Just as example , in photo one of the last proto with full Tango, 999 + NP11303 ( and NP11107 and NP206, in different tests) + EQ-2L + EC8020 and/or EC8010
for my friend, good results. lot of test lab!!
The Freq. answer of that proto
First of all do you have a proper test set?
Only following the results of a test lab you can reach the right results.
If you check the pdf from Docali :
Practical Considerations
on Designing and Measering
LCR RIAA Filters
Dr. Bernd Ahlswede (May 2018)
it gives you the right direction, with a good sound card and a proper setting you will have a good start point.
In addition, check each stage one by one regarding freq. answer and distortion and gain.
The final results is strictly related ot the error on each stage, it is the sum.
Just as example , in photo one of the last proto with full Tango, 999 + NP11303 ( and NP11107 and NP206, in different tests) + EQ-2L + EC8020 and/or EC8010
for my friend, good results. lot of test lab!!
The Freq. answer of that proto
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
Thank you very much, Walge, for posting this test data.walge wrote: ↑Thu Jul 01, 2021 8:04 am to WE
First of all do you have a proper test set?
Only following the results of a test lab you can reach the right results.
If you check the pdf from Docali :
Practical Considerations
on Designing and Measering
LCR RIAA Filters
Dr. Bernd Ahlswede (May 2018)
it gives you the right direction, with a good sound card and a proper setting you will have a good start point.
In addition, check each stage one by one regarding freq. answer and distortion and gain.
The final results is strictly related ot the error on each stage, it is the sum.
Just as example , in photo
Tango_LCR_proto.JPG
one of the last proto with full Tango, 999 + NP11303 ( and NP11107 and NP206, in different tests) + EQ-2L + EC8020 and/or EC8010
for my friend, good results. lot of test lab!!
The Freq. answer of that proto
LCR_freq_resp_full_Tango_8020.jpg
Yes, I have the proper test equipment, etc., to perform this testing.
I do have two questions for you: How did you bias the EC8010 and EC8020 in your LCR phono preamp? LED bias, battery bias, fixed bias, etc.? And what was the Ep and Ip that you used that provided the best operating point for these tubes, in terms of lowest distortion, etc.?
With Best Regards,
WE300B
Re: Step-up after LCR
Hi
regarding bias personally I don't like to mix sand with vacuum![Smile :)](./images/smilies/icon_smile.gif)
So I used a cathode resistor and cap with around 20-25 mA of bias and around 200-220 Vdc perfectly stabilized with ss ( sand!), this because I got a very good s/n that is one of most important goal, in my opinion.
Also the Thd was very fine
Walter
regarding bias personally I don't like to mix sand with vacuum
![Smile :)](./images/smilies/icon_smile.gif)
So I used a cathode resistor and cap with around 20-25 mA of bias and around 200-220 Vdc perfectly stabilized with ss ( sand!), this because I got a very good s/n that is one of most important goal, in my opinion.
Also the Thd was very fine
Walter
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
Hi Walter,walge wrote: ↑Fri Jul 02, 2021 3:19 pm Hi
regarding bias personally I don't like to mix sand with vacuum
So I used a cathode resistor and cap with around 20-25 mA of bias and around 200-220 Vdc perfectly stabilized with ss ( sand!), this because I got a very good s/n that is one of most important goal, in my opinion.
Also the Thd was very fine
Walter
Yes, I agree; I don't usually like to mix vacuum tubes with semiconductors, but I do use the popular Rod Coleman constant current regulators for providing DC heating for my WE-300B SE output tubes, and the 45 driver stage!
Cathode bias is fine by the use of a resistor, and that would be my preferred approach; the issue is of course the very large value electrolytic bypass capacitor that is required. These very high Gm/high-u tubes, such as the D3A, the EC8010, WE-417A, etc., require very low values of Rk, and therefore very large values of bypass capacitor (Ck) to obtain decent low frequency response. Out of curiosity, what value bypass capacitor are you using, and what brand?
Using this approach, I don't doubt that the measured distortion is very low, but I would question the subjective quality of the sound using a large cathode bypass capacitor!
![Very Happy :D](./images/smilies/icon_biggrin.gif)
All the Best,
WE300B
Re: Step-up after LCR
Hi
I have tested some cap for bypass, last productions, declared 105° and 125 °.
They are very fine, in addition with a proper bypass, in my opinion, the final results will be at high level.
I use value around 330-1000 uF ( depends on Rk) , 25 Vdc; Nichicon, Kemet. I have some Cerafine 220 uF, very good and some lasts FK Rubycon
The bypass are in a range of 0,22-1 uF pp ( Wima mainly)
Regarding 300B, please take a look on two my projrct of some years ago
One is a stereo p-p. The other is a monos with 4 300B in p-p Then a special one. Double p-p of 300B for my friend. The box was specially made by a archistar studio
Walter
I have tested some cap for bypass, last productions, declared 105° and 125 °.
They are very fine, in addition with a proper bypass, in my opinion, the final results will be at high level.
I use value around 330-1000 uF ( depends on Rk) , 25 Vdc; Nichicon, Kemet. I have some Cerafine 220 uF, very good and some lasts FK Rubycon
The bypass are in a range of 0,22-1 uF pp ( Wima mainly)
Regarding 300B, please take a look on two my projrct of some years ago
One is a stereo p-p. The other is a monos with 4 300B in p-p Then a special one. Double p-p of 300B for my friend. The box was specially made by a archistar studio
Walter
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
Hello Walter!
I know this is off topic in this thread, but those are very beautiful 300B amplifiers!
Here is a photo of my homebrew 2-stage WE-300B amplifier, driven by a transformer-coupled 245. The power supply for each amplifier has two separate HV supplies, one for the 245 driver, and one for the WE-300B SE output stage. It uses a pair of 83 mercury vapor rectifier tubes; one for the 245, and one for the WE-300B.
I use a transformer volume control (TVC) with this system; the TVC is directly to the right of the amplifier unit.
All the Best,
WE300B
I know this is off topic in this thread, but those are very beautiful 300B amplifiers!
Here is a photo of my homebrew 2-stage WE-300B amplifier, driven by a transformer-coupled 245. The power supply for each amplifier has two separate HV supplies, one for the 245 driver, and one for the WE-300B SE output stage. It uses a pair of 83 mercury vapor rectifier tubes; one for the 245, and one for the WE-300B.
I use a transformer volume control (TVC) with this system; the TVC is directly to the right of the amplifier unit.
All the Best,
WE300B
- Attachments
-
- 45-300B Amplifier With External LC PSU Filter 1.jpg (1.18 MiB) Viewed 127708 times
Re: Step-up after LCR
Very fine!!!
Walter
Walter
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
I did some calculations today with regard to the Miller Effect and the D3A input tube (the first gain stage in this circuit).
The Miller C that I calculated for the D3A in triode connection is about 468 pf, based upon a u+1 of 78.
The load impedance that my moving coil cartridge (EMT TSD-15) likes to see is around 250 ohms. The TSD-15 has a source Z of 25 ohms, and assume the DCR of the primary winding of my MC-SUT to be around 18 ohms. If I use an MC-SUT with a secondary Z of 70,000 ohms, and a primary Z of 250 ohms, the Miller Effect will begin to become something of an issue. The F3 point approaches 23.3 Khz. Not good....
Reducing the secondary Z of the MC-SUT to 50,000 ohms improves the F3 point to around 30 Khz. That is of course much better.
Further reducing the secondary Z of the MC-SUT to 30,000 ohms improves the F3 point to around 54 Khz, which is very workable, and this is no longer an issue with this circuit.
The difference in MC-SUT gain between a secondary Z of 70,000 ohms and 30,000 ohms is around 3.7 dB, which I can easily live with.
The second D3A gain stage in this circuit (after the 600 ohm LCR network) has an F3 point of around 34 Khz, which is workable.
With a high u tube such as the D3A in triode connection, the Miller Effect is clearly not our friend, unless we elect to use the tube in a cascode configuration, which I don't want to do.....
The Miller C that I calculated for the D3A in triode connection is about 468 pf, based upon a u+1 of 78.
The load impedance that my moving coil cartridge (EMT TSD-15) likes to see is around 250 ohms. The TSD-15 has a source Z of 25 ohms, and assume the DCR of the primary winding of my MC-SUT to be around 18 ohms. If I use an MC-SUT with a secondary Z of 70,000 ohms, and a primary Z of 250 ohms, the Miller Effect will begin to become something of an issue. The F3 point approaches 23.3 Khz. Not good....
Reducing the secondary Z of the MC-SUT to 50,000 ohms improves the F3 point to around 30 Khz. That is of course much better.
Further reducing the secondary Z of the MC-SUT to 30,000 ohms improves the F3 point to around 54 Khz, which is very workable, and this is no longer an issue with this circuit.
The difference in MC-SUT gain between a secondary Z of 70,000 ohms and 30,000 ohms is around 3.7 dB, which I can easily live with.
The second D3A gain stage in this circuit (after the 600 ohm LCR network) has an F3 point of around 34 Khz, which is workable.
With a high u tube such as the D3A in triode connection, the Miller Effect is clearly not our friend, unless we elect to use the tube in a cascode configuration, which I don't want to do.....
-
- Posts: 2091
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Re: Step-up after LCR
When you start getting into higher impedance cartridges and SUT's the miller capacitance (and cable capacitance) of the input of your phono becomes an important consideration. In addition to the simple filter you speak of, the additional capacitance pushes the LC resonance down in frequency and depending on the SUT design this can get close to the audio band.
Here is a somewhat abused peerless 4722 showing what happens when a low Z cart is used into a light but varying capacitive load. Granted moving from a 500K load to a 50K load will damp the resonance as will moving from a 4Ω to a 25Ω will effectively eliminate the peak but in the case of the 4722 I would still expect -1dB @ 10kHz from a 500pF load.
For my designs I have set 1:14 as the maximum ratio for the high Z carts (>20Ω) into 47K in parallel with <200p of capacitance. It isn't until the resonant frequency is a decade above the audio band that the simple behavior you describe dominates and then it is important to note the phase characteristics of the capacitive load.
Here is a somewhat abused peerless 4722 showing what happens when a low Z cart is used into a light but varying capacitive load. Granted moving from a 500K load to a 50K load will damp the resonance as will moving from a 4Ω to a 25Ω will effectively eliminate the peak but in the case of the 4722 I would still expect -1dB @ 10kHz from a 500pF load.
For my designs I have set 1:14 as the maximum ratio for the high Z carts (>20Ω) into 47K in parallel with <200p of capacitance. It isn't until the resonant frequency is a decade above the audio band that the simple behavior you describe dominates and then it is important to note the phase characteristics of the capacitive load.
Get Your Fix
www.hifiheroin.com
www.hifiheroin.com
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
Hi Dave,
Thanks very much for your comments and guidance, and for taking and posting this data. It is very helpful!
I finished wiring up my new LCR phono preamp last night. Today, I'll start wiring up the power supply for it.
I'm looking forward to putting it on the bench and taking data. Should be interesting....the big variable in my mind is the correct implementation of the 600 ohm termination impedance for the LCR networks.
All the Best,
WE300B
Thanks very much for your comments and guidance, and for taking and posting this data. It is very helpful!
I finished wiring up my new LCR phono preamp last night. Today, I'll start wiring up the power supply for it.
I'm looking forward to putting it on the bench and taking data. Should be interesting....the big variable in my mind is the correct implementation of the 600 ohm termination impedance for the LCR networks.
All the Best,
WE300B
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
I have read all of the comments and reviewed the theoretical and actual measured data here. I believe all of this data is based upon a simulated/modeled analysis of the predicted amplitude and phase response of strictly the LCR network itself, as well as a possible circuit using a 600 ohm LCR network.
The circuit posted earlier in this post by Stegie, with the Lundahl LL1692As plate-to-line and Sowter 9063 step-up transformers is very interesting and elegant, and it is presented as a possible basis for an actual phono preamplifier using a 600 ohm LCR network. The link to Dr. Ahlswede's (Docali) white paper on the subject of designing and measuring LCR RIAA networks is particularly interesting and informative. Although Dr. Ahlswede did take data on his LCR1200 network when properly terminated and looking into a Sowter 1:4 600/10K step-up transformer, and the data was very encouraging, he did not provide the measured data of a complete RIAA preamp with his LCR1200 network and these Lundahl plate-to-line and other elements in the signal path.
My question is this; it is indeed possible to obtain +/- 0.25 dB (or similar) accuracy over the entire RIAA curve with a properly designed and terminated all-transformer coupled LCR network? When a 600 ohm LCR network is used in the context of the complete circuit posted here, with three transformers in the signal path, and two vacuum tubes, what is the actual measured (not modeled) amplitude and phase accuracy of the RIAA curve?
I question if it is possible for such a preamplifier circuit, with all of the complex (transformers and LCR network) impedances within the signal path, to actually adhere to +/- 0.25 dB accuracy of the RIAA playback curve, although commercially-made "high-end" vacuum tube preamps that don't utilize transformer coupling and use passive or active RC networks for establishing the playback curve, claim to obtain +/-0.5 dB accuracy or better, which I don't doubt.
Has anyone built this three transformer/two-tube LCR preamp, or a similar circuit, and measured the actual results with an inverse RIAA network? It may well be another example of a circuit that indeed sounds spectacular, but measures poorly in terms of it's RIAA accuracy over the 20 to 20 Khz range.
Of course, several commercial and very highly sonically regarded examples of a somewhat similar (2-tube/transformer coupled, with LCR equalization) circuit exist from a well-known German designer, but has anyone measured the RIAA accuracy of these units? That would be most interesting data to see.
Comments on this are very much welcome!
Best Regards,
WE300B
The circuit posted earlier in this post by Stegie, with the Lundahl LL1692As plate-to-line and Sowter 9063 step-up transformers is very interesting and elegant, and it is presented as a possible basis for an actual phono preamplifier using a 600 ohm LCR network. The link to Dr. Ahlswede's (Docali) white paper on the subject of designing and measuring LCR RIAA networks is particularly interesting and informative. Although Dr. Ahlswede did take data on his LCR1200 network when properly terminated and looking into a Sowter 1:4 600/10K step-up transformer, and the data was very encouraging, he did not provide the measured data of a complete RIAA preamp with his LCR1200 network and these Lundahl plate-to-line and other elements in the signal path.
My question is this; it is indeed possible to obtain +/- 0.25 dB (or similar) accuracy over the entire RIAA curve with a properly designed and terminated all-transformer coupled LCR network? When a 600 ohm LCR network is used in the context of the complete circuit posted here, with three transformers in the signal path, and two vacuum tubes, what is the actual measured (not modeled) amplitude and phase accuracy of the RIAA curve?
I question if it is possible for such a preamplifier circuit, with all of the complex (transformers and LCR network) impedances within the signal path, to actually adhere to +/- 0.25 dB accuracy of the RIAA playback curve, although commercially-made "high-end" vacuum tube preamps that don't utilize transformer coupling and use passive or active RC networks for establishing the playback curve, claim to obtain +/-0.5 dB accuracy or better, which I don't doubt.
Has anyone built this three transformer/two-tube LCR preamp, or a similar circuit, and measured the actual results with an inverse RIAA network? It may well be another example of a circuit that indeed sounds spectacular, but measures poorly in terms of it's RIAA accuracy over the 20 to 20 Khz range.
Of course, several commercial and very highly sonically regarded examples of a somewhat similar (2-tube/transformer coupled, with LCR equalization) circuit exist from a well-known German designer, but has anyone measured the RIAA accuracy of these units? That would be most interesting data to see.
Comments on this are very much welcome!
Best Regards,
WE300B
Last edited by westernelectric300b on Wed Sep 29, 2021 4:45 pm, edited 1 time in total.
-
- Posts: 2091
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Re: Step-up after LCR
If you consider the basics, the LCR has its irregularities in the middle of the audio band where transformers are flat so the two should not effect each other. Granted the Transformers will have bandwidth limiting issues but that is independent of the type of EQ used. I guess I am trying to say is that if each part of the circuit behaves acceptably without the other, combining then should not cause any other issues.
dave
dave
Get Your Fix
www.hifiheroin.com
www.hifiheroin.com
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
Dave, I agree with that.
I think any amplitude anomalies, particularly premature roll-off, would occur in the low (<100 Hz) and high (>10 Khz) portions of the audio band, due in large part to the cascaded complex impedances such a circuit would probably exhibit.
But we are just speculating here. As I had suggested in my post, I'd like to see the actual measured amplitude response over the 20 to 20 Khz band of a 2 tube/3-transformer/600 ohm LCR open-loop phono preamp such as the circuit or circuits described in this post, or perhaps the product offered by the well-known designer in Germany of that very highly regarded product.
Until otherwise, I suspect this kind of circuit may suffer from the kind of amplitude response issues described above. And although these circuits probably sound fantastic, which of course is the most important issue, this may well be another case of the circuit sounding amazing, but measuring poorly, at least with regard to bandwidth and flatness.
Best Regards,
WE300B
I think any amplitude anomalies, particularly premature roll-off, would occur in the low (<100 Hz) and high (>10 Khz) portions of the audio band, due in large part to the cascaded complex impedances such a circuit would probably exhibit.
But we are just speculating here. As I had suggested in my post, I'd like to see the actual measured amplitude response over the 20 to 20 Khz band of a 2 tube/3-transformer/600 ohm LCR open-loop phono preamp such as the circuit or circuits described in this post, or perhaps the product offered by the well-known designer in Germany of that very highly regarded product.
Until otherwise, I suspect this kind of circuit may suffer from the kind of amplitude response issues described above. And although these circuits probably sound fantastic, which of course is the most important issue, this may well be another case of the circuit sounding amazing, but measuring poorly, at least with regard to bandwidth and flatness.
Best Regards,
WE300B
Re: Step-up after LCR
Hi
In attach the proto of Full Tango with 999 in input, interstage NP206, EQ2L, NP206.
I have also tested 11303 and 11107, rares.
With EC8020; then check with EC8010 and E280F
In addition the freq. answer of the phono with some adjustement on components.
The results is very fine but to get this you must have a proper test setup.
My anti-riaa comes from AP system 2 and the curve is done by software so it is perfect even I change the Zout of generator.
In every case some little changes must be done on resistors and cap of the network.
Walter
In attach the proto of Full Tango with 999 in input, interstage NP206, EQ2L, NP206.
I have also tested 11303 and 11107, rares.
With EC8020; then check with EC8010 and E280F
In addition the freq. answer of the phono with some adjustement on components.
The results is very fine but to get this you must have a proper test setup.
My anti-riaa comes from AP system 2 and the curve is done by software so it is perfect even I change the Zout of generator.
In every case some little changes must be done on resistors and cap of the network.
Walter
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
Hi Walter,
Thanks very much for posting actual test data of an LCR preamp with two tubes and three audio transformers in the signal path. That is indeed very encouraging.
Can you clarify your comment about, "you must have a proper test setup"? I understand that you are using an Audio Precision analyzer. For checking the accuracy of my phono stages, I use an H/P Model 652A audio generator, a Hagerman inverse RIAA network, and an H/P Model 333A distortion analyzer/audio voltmeter. The audio voltmeter is calibrated in volts and in dB; setting the audio generator output to 1000 hz, I set the reference level on the audio voltmeter to 0 dB. Any deviation from the ideal RIAA curve can then be read directly off of the dB scale of the meter.
I have used this set-up for years for checking the RIAA accuracy of my phono preamps. Although this set-up may not be as fast or have the kind of resolution as the Audio Precision equipment, it can still provide me with measurement resolution to 0.2 dB over the range of 20 to 20 Khz, or wider if I desire.
When you mentioned, "the freq. answer of the phono with some adjustment on components", what components are you referring to in order to obtain these adjustments? It obviously can't be the components within the Tango EQP-600 network, as the L, R, & C values are fixed in that network, and cannot be changed.
Best Regards,
WE300B
Thanks very much for posting actual test data of an LCR preamp with two tubes and three audio transformers in the signal path. That is indeed very encouraging.
Can you clarify your comment about, "you must have a proper test setup"? I understand that you are using an Audio Precision analyzer. For checking the accuracy of my phono stages, I use an H/P Model 652A audio generator, a Hagerman inverse RIAA network, and an H/P Model 333A distortion analyzer/audio voltmeter. The audio voltmeter is calibrated in volts and in dB; setting the audio generator output to 1000 hz, I set the reference level on the audio voltmeter to 0 dB. Any deviation from the ideal RIAA curve can then be read directly off of the dB scale of the meter.
I have used this set-up for years for checking the RIAA accuracy of my phono preamps. Although this set-up may not be as fast or have the kind of resolution as the Audio Precision equipment, it can still provide me with measurement resolution to 0.2 dB over the range of 20 to 20 Khz, or wider if I desire.
When you mentioned, "the freq. answer of the phono with some adjustment on components", what components are you referring to in order to obtain these adjustments? It obviously can't be the components within the Tango EQP-600 network, as the L, R, & C values are fixed in that network, and cannot be changed.
Best Regards,
WE300B
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
Looking at the schematic diagram of the 2-stage transformer-coupled LCR phono preamp posted here earlier, I see a potential issue.
The LCR network is a T-network, and it's input and output impedances are identical. We also know that the network must either be terminated (as in the case of this circuit) in a 600 ohm load impedance, in which case the source or driving impedance can be essentially anything other than 600 ohms, with a source impedance substantially less than 600 ohms being more desirable. Or the network can be driven from a 600 ohm source, in which case the terminating impedance must be some value other than 600 ohms, and ideally substantially higher than 600 ohms. Otherwise, serious RIAA accuracy issues will occur.
The schematic indicates a load or terminating impedance of 600 ohms at the output of the LCR network, which is the reflected impedance as it appears at the primary of the Sowter 9063 step-up transformer. The alternative, which I prefer, is to terminate the LCR network in a 600 ohm resistive load, and then either terminate the secondary of the step-up transformer with some value such as 50K or so, or leave the secondary unterminated.
So the real issue now becomes the source or driving impedance to the LCR network. As I see it, the secondary of the LL1692A plate-to-line transformer is providing something on the order of a 600 ohm driving impedance to the network, which is undesirable for driving the network when it is terminated in a 600 ohm load impedance. One viable solution would entail changing the transformer to one having a secondary impedance of 150 ohms to drive the input to the network, with some loss of gain to the overall circuit. The other would be to perhaps retain the Lundahl LL1692A transformer with it's 600 ohm secondary, and insert a resistor in series with the input to the LCR network, so as to increase the apparent source impedance seen by the input to the LCR network.
What are the thoughts on the solution of inserting a resistor in series between the LL1692A secondary and the input to the LCR network?
WE300B
The LCR network is a T-network, and it's input and output impedances are identical. We also know that the network must either be terminated (as in the case of this circuit) in a 600 ohm load impedance, in which case the source or driving impedance can be essentially anything other than 600 ohms, with a source impedance substantially less than 600 ohms being more desirable. Or the network can be driven from a 600 ohm source, in which case the terminating impedance must be some value other than 600 ohms, and ideally substantially higher than 600 ohms. Otherwise, serious RIAA accuracy issues will occur.
The schematic indicates a load or terminating impedance of 600 ohms at the output of the LCR network, which is the reflected impedance as it appears at the primary of the Sowter 9063 step-up transformer. The alternative, which I prefer, is to terminate the LCR network in a 600 ohm resistive load, and then either terminate the secondary of the step-up transformer with some value such as 50K or so, or leave the secondary unterminated.
So the real issue now becomes the source or driving impedance to the LCR network. As I see it, the secondary of the LL1692A plate-to-line transformer is providing something on the order of a 600 ohm driving impedance to the network, which is undesirable for driving the network when it is terminated in a 600 ohm load impedance. One viable solution would entail changing the transformer to one having a secondary impedance of 150 ohms to drive the input to the network, with some loss of gain to the overall circuit. The other would be to perhaps retain the Lundahl LL1692A transformer with it's 600 ohm secondary, and insert a resistor in series with the input to the LCR network, so as to increase the apparent source impedance seen by the input to the LCR network.
What are the thoughts on the solution of inserting a resistor in series between the LL1692A secondary and the input to the LCR network?
WE300B
Re: Step-up after LCR
What you state here is mathematically not correct. For a correct filter response both sides of the t-network must be "terminated" with 600 ohms. I showed the effect of different driving impedance in my wp. Everything else than a driving impedance of 600 ohms is less correct. Then we come to real inductors and the real network (and the amp circuit) and you have to measure how good your result is. Practically a driving impedance of 150 ohms works quite good. To use an additional series resistor before the LCR is absolutely correct if you want a better LF eq result.
I again recommend the study of Claus' excellent work.
There was a comment that I did not show results of a complete amp circuit. This is correct. Spare time is limited and the project never came alive but with the appropriate inductors this would work good. In the meantime Mr. Sowter sold his business so I am not sure if the new management would continue to serve such "uncommon" requests like my for 1k2 lcr inductors.
I again recommend the study of Claus' excellent work.
There was a comment that I did not show results of a complete amp circuit. This is correct. Spare time is limited and the project never came alive but with the appropriate inductors this would work good. In the meantime Mr. Sowter sold his business so I am not sure if the new management would continue to serve such "uncommon" requests like my for 1k2 lcr inductors.
Re: Step-up after LCR
to western
I mean that you may trim the perfect curve puttin in parallel to the 600 ohm on the output of Lcr network a proper resistors.
In my case with the source driving of around 200 ohm ( CF of the PCC88) I put 1k8 on parallel of 600 ohm.
The response without this trimming is little bit worst at 20 Hz and still good at 20kHz; in my opinion to reach the perfect cutve is always better and it is my goal.
The total gain with a selected tubes is 42 dB.
I have seen little differences also changing the CF tube brand (but always family of 88)
About antiriaa there is a beautiful document from Docali where is described the setting with ARTA sw to obtain a curve without a possible issue of the Zsource that you can have with a passive network.
In attach the last proto with full dual mono LCR board.
I have had a beautiful reports from someone ( not so many!!
) people have built my schematic
Walter
I mean that you may trim the perfect curve puttin in parallel to the 600 ohm on the output of Lcr network a proper resistors.
In my case with the source driving of around 200 ohm ( CF of the PCC88) I put 1k8 on parallel of 600 ohm.
The response without this trimming is little bit worst at 20 Hz and still good at 20kHz; in my opinion to reach the perfect cutve is always better and it is my goal.
The total gain with a selected tubes is 42 dB.
I have seen little differences also changing the CF tube brand (but always family of 88)
About antiriaa there is a beautiful document from Docali where is described the setting with ARTA sw to obtain a curve without a possible issue of the Zsource that you can have with a passive network.
In attach the last proto with full dual mono LCR board.
I have had a beautiful reports from someone ( not so many!!
![Smile :)](./images/smilies/icon_smile.gif)
Walter
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
Hello Docali, and thank you very much for your reply.docali wrote: ↑Thu Oct 07, 2021 8:41 am What you state here is mathematically not correct. For a correct filter response both sides of the t-network must be "terminated" with 600 ohms. I showed the effect of different driving impedance in my wp. Everything else than a driving impedance of 600 ohms is less correct. Then we come to real inductors and the real network (and the amp circuit) and you have to measure how good your result is. Practically a driving impedance of 150 ohms works quite good. To use an additional series resistor before the LCR is absolutely correct if you want a better LF eq result.
I again recommend the study of Claus' excellent work.
There was a comment that I did not show results of a complete amp circuit. This is correct. Spare time is limited and the project never came alive but with the appropriate inductors this would work good. In the meantime Mr. Sowter sold his business so I am not sure if the new management would continue to serve such "uncommon" requests like my for 1k2 lcr inductors.
Can you clarify for me several of your points?:
1) "For a correct filter response both sides of the t-network must be "terminated" with 600 ohms. I showed the effect of different driving impedance in my wp. Everything else than a driving impedance of 600 ohms is less correct." I have read your excellent white paper on the subject several times. Are you suggesting here that for the maximum RIAA accuracy, the LCR network must be driven by a 600 ohm source impedance? You also state that a driving impedance of 150 ohms is quite good. We both agree that the LCR network absolutely must be terminated in 600 ohms, so the only real variable here in terms of the RIAA accuracy of the LCR network is the source impedance driving the network. Which is correct, as I am a little confused? I thought that if the LCR network was terminated in a 600 ohm resistive load, the source impedance was then largely not critical.
2) Can you provide us with the link to Claus's paper on the subject?
3) The first step-down transformer (the one driving the LCR network) shown in the complete preamp circuit in this post, is a Lundahl LL1692A, with a 4:1.75 step-down turns ratio. Would a Lundahl LL1692A with a 3.5:1 step-down turns ratio in the Alt Q configuration provide a better (or lower) impedance match to the 600 ohm LCR network?
Thanks & Regards,
WE300B
-
- Posts: 2091
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Re: Step-up after LCR
Taken out of context of the entire paper, it is unclear what docali is referring to in this statement.
If you look at the response of the simulations in the amplitude realm you will see the same frequency response for each and the only hting that changes appreciably is the overall gain which is simply based on the Rsource to Rload ratio. The picture below shows the relative results from the tango network driven by 50, 600 and 1200Ω. I did not bother with the 1 meg source since effectively shuts the circuit off.
Docali is specifically talking about the LCR network as source impedance driving the primary of a transformer and when considering behavior below 100hz where this "driving impedance" varies dropping below 600hz can only help to extend the low frequency based on the finite inductance it works against. Driving a "stock" tango from below 600Ω and terminating it with an inductance like the primary of a transformer will net the sub 100hz anomalies that docali refers to but they can easily be addresses by small changes in the LCR part values. For a 50Ω source and a 600Ω || 20hy termination the 1.8Hy inductor needs to be ~2.1hy and the 495Ω parallel resistors need to be ~525Ω.
So to qualify the original statement taken out of context I would change it to:
For a correct filter response when terminated with an inductive load both sides of the Tango /pultec LCR t-network must be "terminated" with 600 ohms.
![Smile :-)](./images/smilies/icon_smile.gif)
dave
In the amplitude realm if the output is terminated with 600Ω, a change from a input termination (source impedance) from 150Ω to 600Ω has a trivial effect across the frequency range in question. What his graph shows is how a varying drive impedance forces the filter to deviate from the "ideal" T-network. drive it from lower than 600Ω and the networks output impedance drops below 100hz. drive it form above 600Ω and the output Z of the network goes up.For a correct filter response both sides of the t-network must be "terminated" with 600 ohms
If you look at the response of the simulations in the amplitude realm you will see the same frequency response for each and the only hting that changes appreciably is the overall gain which is simply based on the Rsource to Rload ratio. The picture below shows the relative results from the tango network driven by 50, 600 and 1200Ω. I did not bother with the 1 meg source since effectively shuts the circuit off.
Docali is specifically talking about the LCR network as source impedance driving the primary of a transformer and when considering behavior below 100hz where this "driving impedance" varies dropping below 600hz can only help to extend the low frequency based on the finite inductance it works against. Driving a "stock" tango from below 600Ω and terminating it with an inductance like the primary of a transformer will net the sub 100hz anomalies that docali refers to but they can easily be addresses by small changes in the LCR part values. For a 50Ω source and a 600Ω || 20hy termination the 1.8Hy inductor needs to be ~2.1hy and the 495Ω parallel resistors need to be ~525Ω.
So to qualify the original statement taken out of context I would change it to:
For a correct filter response when terminated with an inductive load both sides of the Tango /pultec LCR t-network must be "terminated" with 600 ohms.
Oh this irony here... you guys are having this discussion on the website of someone who winds transformers for a living and for the past 15 years on this very forum advocated for and wound inductors for LCR circuits at other than 600Ω.Mr. Sowter sold his business so I am not sure if the new management would continue to serve such "uncommon" requests like my for 1k2 lcr inductors.
![Smile :-)](./images/smilies/icon_smile.gif)
dave
Get Your Fix
www.hifiheroin.com
www.hifiheroin.com
-
- Posts: 28
- Joined: Fri Apr 30, 2021 1:29 am
Re: Step-up after LCR
Hi Dave,
In the immortal words of the classic film, "Cool Hand Luke", I think what we have here is a failure to communicate!
Per your summation in your last post: "So to qualify the original statement taken out of context I would change it to:
For a correct filter response when terminated with an inductive load both sides of the Tango /pultec LCR t-network must be "terminated" with 600 ohms".
We all know that for the optimum amplitude response/flatness of the RIAA reproduction curve from the LCR network, the network must be terminated in a 600 ohm load. That is a given. A 600 ohm resistor at the output of the LCR network does the job nicely here.
However, can you clarify your statement about the input of the network? I think this is where the confusion arises. Assume the 600 ohm LCR network is being driven by a typical plate-to-line transformer having some relatively low secondary impedance, anywhere from 150 to 600 ohms or so, such as the Lundahl 1692A. Are you suggesting here that the secondary of the driving transformer must be terminated in a 600 ohm resistor? This is where I am confused. If the LCR network is a 600 ohm device, doesn't it present a 600 ohm load impedance to the secondary of the transformer driving the network, thereby eliminating the need for the 600 ohm resistor across the secondary of the plate-to-line transformer, and as such, across the input to the LCR network?
Thanks!
WE300B
In the immortal words of the classic film, "Cool Hand Luke", I think what we have here is a failure to communicate!
Per your summation in your last post: "So to qualify the original statement taken out of context I would change it to:
For a correct filter response when terminated with an inductive load both sides of the Tango /pultec LCR t-network must be "terminated" with 600 ohms".
We all know that for the optimum amplitude response/flatness of the RIAA reproduction curve from the LCR network, the network must be terminated in a 600 ohm load. That is a given. A 600 ohm resistor at the output of the LCR network does the job nicely here.
However, can you clarify your statement about the input of the network? I think this is where the confusion arises. Assume the 600 ohm LCR network is being driven by a typical plate-to-line transformer having some relatively low secondary impedance, anywhere from 150 to 600 ohms or so, such as the Lundahl 1692A. Are you suggesting here that the secondary of the driving transformer must be terminated in a 600 ohm resistor? This is where I am confused. If the LCR network is a 600 ohm device, doesn't it present a 600 ohm load impedance to the secondary of the transformer driving the network, thereby eliminating the need for the 600 ohm resistor across the secondary of the plate-to-line transformer, and as such, across the input to the LCR network?
Thanks!
WE300B
-
- Posts: 2091
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Re: Step-up after LCR
There are a couple of things going on here. When you terminate the "stock" LCR with both 600Ω and an inductance as is happening in this thread, the source Z driving it must also be 600Ω or things shift. The attached picture shows a stock tango driven from 150Ω (green) and 600Ω (blue) when loaded with a 20hy inductance. Granted a 0.6dB drop @ 20hz is not the end of the world and it is also important to note that the "droop" in the 150Ω source Z case can be corrected for in a self built LCT network by adjusting a couple of the values.
dave
dave
Get Your Fix
www.hifiheroin.com
www.hifiheroin.com
Re: Step-up after LCR
@Dave: Sorry for my last (confused) Emails. I should have shut up and measured first.
To everyone of the smart people discussing LCR filters here: I built a phono pre:
Triode -> 4:1 interstage -> 600R LCR -> 1:4 step up -> triode -> 1:1 output.
Dave wound the 4:1 interstage and in circuit (measured at the input of the LCR) it is flat to 90kHz.
Here is a measurement of the output of the LCR. Red trace: output attached to probe/measurement gear, otherwise open. Blue trace: output attached to the 1:4 stepup (Lundahl 7903). Sorry for the noise around 40k, you still get what's going on...)
I am not happy with the frequency response; the blue trace is pretty much what I get at the output of the preamp.
When I got the discussion right, so far it was about lf roll off, but not what happens to the hf. Docali writes in his paper "Who wants to further improve these results could bypass the whole LCR module with an RC combination. A 180k resistor in series with 150pF is a good starting point." (p. 13).
Would this be a proper step to get the output closer to the blue trace?
Or is the way to go to change the 4:1 to a 1:1, get a 1k8 LCR and get rid of the 1:4 step up?
@walge : your measured response looks pretty good. Did you have an issue like me along the way?
To everyone of the smart people discussing LCR filters here: I built a phono pre:
Triode -> 4:1 interstage -> 600R LCR -> 1:4 step up -> triode -> 1:1 output.
Dave wound the 4:1 interstage and in circuit (measured at the input of the LCR) it is flat to 90kHz.
Here is a measurement of the output of the LCR. Red trace: output attached to probe/measurement gear, otherwise open. Blue trace: output attached to the 1:4 stepup (Lundahl 7903). Sorry for the noise around 40k, you still get what's going on...)
I am not happy with the frequency response; the blue trace is pretty much what I get at the output of the preamp.
When I got the discussion right, so far it was about lf roll off, but not what happens to the hf. Docali writes in his paper "Who wants to further improve these results could bypass the whole LCR module with an RC combination. A 180k resistor in series with 150pF is a good starting point." (p. 13).
Would this be a proper step to get the output closer to the blue trace?
Or is the way to go to change the 4:1 to a 1:1, get a 1k8 LCR and get rid of the 1:4 step up?
@walge : your measured response looks pretty good. Did you have an issue like me along the way?
-
- Posts: 2091
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Re: Step-up after LCR
V1 / V2 are voltage sources and serve as an easy way to set the tube bias. I have used LED's in this position and they work well. I simply put the expected voltage drop of the LED as a voltage source under the cathode rather than trying to find or model the exact LED I use.
dave
dave
Get Your Fix
www.hifiheroin.com
www.hifiheroin.com
Re: Step-up after LCR
Daniel wrote: ↑Sat Sep 02, 2023 8:24 pm @Dave: Sorry for my last (confused) Emails. I should have shut up and measured first.
To everyone of the smart people discussing LCR filters here: I built a phono pre:
Triode -> 4:1 interstage -> 600R LCR -> 1:4 step up -> triode -> 1:1 output.
Dave wound the 4:1 interstage and in circuit (measured at the input of the LCR) it is flat to 90kHz.
Here is a measurement of the output of the LCR. Red trace: output attached to probe/measurement gear, otherwise open. Blue trace: output attached to the 1:4 stepup (Lundahl 7903). Sorry for the noise around 40k, you still get what's going on...)
LCR Vergleich.jpg
I am not happy with the frequency response; the blue trace is pretty much what I get at the output of the preamp.
When I got the discussion right, so far it was about lf roll off, but not what happens to the hf. Docali writes in his paper "Who wants to further improve these results could bypass the whole LCR module with an RC combination. A 180k resistor in series with 150pF is a good starting point." (p. 13).
Would this be a proper step to get the output closer to the blue trace?
Or is the way to go to change the 4:1 to a 1:1, get a 1k8 LCR and get rid of the 1:4 step up?
@walge : your measured response looks pretty good. Did you have an issue like me along the way?
Hi
but can you post the entire circuit with value?
Walter
Re: Step-up after LCR
Just to loook.
In attach the last stuff I made
LCR "full trafo" 14 irons, all custom made for me.
6 are inside active box.
In attach the last stuff I made
LCR "full trafo" 14 irons, all custom made for me.
6 are inside active box.