Step-up after LCR

the road not taken.
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

After finished a LCR with pcb I will prepare soon another LCR but full trafo.
If you want youn can see something interesting in nexth one moth

Walter
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

V! and V2 are just generic voltages and van be replaced by any biasing method that will give that voltage. I have used LED's in that position and like it. you will have to try a bunch of different types since they all sound and bias up differently.

I think an unbypassed Rk is going to seriously effect the specs required from the transformer but it doesn't hurt to try.

dave
Get Your Fix
www.hifiheroin.com
westernelectric300b
Posts: 28
Joined: Fri Apr 30, 2021 1:29 am

Post by westernelectric300b »

walge wrote:After finished a LCR with pcb I will prepare soon another LCR but full trafo.
If you want youn can see something interesting in nexth one moth

Walter
Hi Walter,

Yes, I would very much like to see your other LCR phono preamp next month.

The design I plan on building is largely modeled around the schematic that was posted here, and will also be utilizing only transformer coupling; no coupling capacitors. I had designed my LCR phono preamp a few years ago, but I never built it, and my design is surprisingly similar to the design in this post. I am currently fabricating the chassis for it.

I plan on using LED bias on the cathodes of the D3As, and audio chokes from Lundahl to isolate the two gain stages in each channel. I would have liked to use the WE Ultrapath design for each gain stage, but as I had mentioned earlier, and with the high mu of the D3As, I'm sure that power supply ripple and the resulting hum modulation at the output of the phono preamp will be a problem. Ultrapath is better suited to low-mu tubes in my experience, and as such, I think the LED bias for the D3As is the best approach here. I don't want to use cathode bypass capacitors, as I do not like the effect they have on the sound.

Could you share the transformer line-up of your current design? I have had good success with Lundahl LL1697A plate-to-line transformers in the past, but I have never used Sowter magnetics. I will be using their 9063 line-to-grid transformers for the step-ups after the LCR EQ modules, as I think they are well suited to that application. As an alternative, I have a pair of Tango NP-8 plate-to-line transformers, and I may consider using them in place of the Sowters. The NP-8s have an impedance ratio of 10K to 600 Ohms, so they would work well here as well as step-up transformers.

I will also be using a pair of the S & B LCR EQ modules that I had purchased from S & B a number of years ago. They are 600 ohm units, and are modeled around the Tango EQP-600 design.
Last edited by westernelectric300b on Sat Oct 02, 2021 2:54 am, edited 4 times in total.
westernelectric300b
Posts: 28
Joined: Fri Apr 30, 2021 1:29 am

Post by westernelectric300b »

dave slagle wrote:V! and V2 are just generic voltages and van be replaced by any biasing method that will give that voltage. I have used LED's in that position and like it. you will have to try a bunch of different types since they all sound and bias up differently.

I think an unbypassed Rk is going to seriously effect the specs required from the transformer but it doesn't hurt to try.

dave
Hi Dave,

Thanks very much for the reply.

I have calculated the Rp for the D3A with an unbypassed Rk, and the Rp rises subtantially to around 8K. This would necessitate a primary impedance of at least 2X Rp for the plate-to-line transformer. The LL1692A would be unworkable here, as it's primary Z is around 7.3K ohms when terminated in a 600 ohm load. The primary impedance of the popular Lundahl LL1660 would also be too low, and of course as the impedance ratio between the primary and the secondary windings increases, the stage gain decreases.

If I could get by with an unbypassed Rk, I would. The current feedback at that node in the circuit would be a good thing.

As such, and for the first iteration of the circuit, I'm going to stick with the LED bias approach, unless someone has a better suggestion.
Last edited by westernelectric300b on Tue Jun 01, 2021 4:50 am, edited 1 time in total.
pecci
Posts: 4
Joined: Sun Jan 03, 2021 6:57 pm

Post by pecci »

Hi, led biasing is fine, and very probably better than cathode bias with resistor and cap, but even better solution is to go with some kind of semi fixed bias, like Thorsten use in Toccata MK3 preamp with E810F tubes.
Actually that way of biasing has a long history and many versions are in Neumann famous tube mics from ’40 and ’50 (U47, U67, M49)
I suppose that Thomas use similar biasing method, by low impedance voltage divider from plate or ht power supply (or both).
For sure precondition for that biasing is to have extremely good filtered supply and use secondaries of mc sut before first tube and of step-up before second tube.
Additionally good thing will be to go with floating and elevated gnd, according to Telefunken TDS for D3a (+Ubg1 = 10V)
My breadboard with LL19933-D3a(T)-LL1692a-LCR600-9063-D3a-LL1692a will be ready in next few months but bias circuit still remain unknown territory for me.
But that circuit is really crucial in this story. Any suggestions and experiences are welcome.
Regards,
westernelectric300b
Posts: 28
Joined: Fri Apr 30, 2021 1:29 am

Post by westernelectric300b »

pecci wrote:Hi, led biasing is fine, and very probably better than cathode bias with resistor and cap, but even better solution is to go with some kind of semi fixed bias, like Thorsten use in Toccata MK3 preamp with E810F tubes.
Actually that way of biasing has a long history and many versions are in Neumann famous tube mics from ’40 and ’50 (U47, U67, M49)
I suppose that Thomas use similar biasing method, by low impedance voltage divider from plate or ht power supply (or both).
For sure precondition for that biasing is to have extremely good filtered supply and use secondaries of mc sut before first tube and of step-up before second tube.
Additionally good thing will be to go with floating and elevated gnd, according to Telefunken TDS for D3a (+Ubg1 = 10V)
My breadboard with LL19933-D3a(T)-LL1692a-LCR600-9063-D3a-LL1692a will be ready in next few months but bias circuit still remain unknown territory for me.
But that circuit is really crucial in this story. Any suggestions and experiences are welcome.
Regards,
Hello Pecci, and thank you for your response and guidance.

Your transformer line-up is identical to mine, except that I will be using a pair of Peerless 4665s as the MC-SUTs in my design. I have used these transformers previously, and with great success.

I am not familiar with the concept of "semi-fixed biasing" that you have described. Can you send me a link to this, so I can study it further? Is it possibly similar to battery bias?

Yes, I agree in that LED biasing is better than conventional cathode bias with a bypass capacitor. I don't use cathode bias with a bypass capacitor in any of my audio designs. LED bias may be very well suited to the D3A, as the tube draws around 20 Ma in operation in my circuit, and this current is ideal for the most linear operation of the LED.

My operating points for the D3A are 140 VDC Eb, and 20 Ma Ib. I actually designed my version of this LCR preamp a number of years ago, and my electrical design and the design posted here are nearly identical, which I find reassuring, in that someone else has independently come to the same electrical design as I did!

I am using the Stevens and Billington 600 ohm LCR networks, which are no longer available.

The sheet metal for my LCR phono preamp and the external power supply unit are close to completion. The delay in completing this project are the long deliveries from Lundahl for the LL1668 isolating/decoupling chokes, the LL1692As, and the 9063s from Sowter.

I look forward to corresponding with you and anyone else on this most interesting project.

All the Best!
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

This is the las version of my LCR
I am playing now with the version full trafos, (possibly) two stage.

Walter
Attachments
LCR_Phono_pcb.JPG
LCR_Phono_pcb.JPG (87.6 KiB) Viewed 77296 times
pecci
Posts: 4
Joined: Sun Jan 03, 2021 6:57 pm

Post by pecci »

Hello WE300B,
I attach schematic of Thorsten LCR. You can find their full description of that preamp on Jogis Roehrenbunde site, but in German only. Please note SUT on input side missing in schematic and Thorsten use S&B TX103.
Additionally, I add one schematic from Neumann, just for illustration of that biasing concept.
I don’t know what is the correct name of that bias but that direction looks very promising for me and I will be deeply grateful for any help on that road.
Key point here is to biasing tube (especially high Gm types like D3a) on elastic and floating way by mutual interaction of voltage dividers from plate and ht supply to achieve better musicality with more relaxed sound, and partly avoid painful tube matching.
Your OP for D3a is close to my, maybe I prefer slightly lower Ib and Ub for about 10%.
And yes, lead time from transformer producer become longer and longer. I was waiting 9063 for almost 4 months. In comparison to them Dave is clear champion. Thanks Dave.
What is your suggestion regarding LED type?
Best regards,
Attachments
Phono Equaliser Schematic.gif
Phono Equaliser Schematic.gif (13.91 KiB) Viewed 77292 times
Phono Equaliser Powersupply Schematic.gif
Phono Equaliser Powersupply Schematic.gif (12.87 KiB) Viewed 77292 times
Neumann M49.jpg
Neumann M49.jpg (244.92 KiB) Viewed 77292 times
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

The schematic on Thorsten is not correct.
In attach the one correct

The point of view of Th is that the flat response on RIAA is not so important; this is an opinion, of course.
The solution he did is about the less precise configuration from Tango EQ600 data sheet.

Walter
Attachments
Figura_10_Phono_Thorsten.jpg
Figura_10_Phono_Thorsten.jpg (116.86 KiB) Viewed 77288 times
pecci
Posts: 4
Joined: Sun Jan 03, 2021 6:57 pm

Post by pecci »

Hi Walter,
Yes that schematic is problematic in few points, but for me is very interesting regarding biasing.
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

Regaridng the bias the last solution, cathode to ground, is very good, it depends of the tube.

W
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

I like and use the pentode as the first stage but the use of a 600Ω load which appears in parallel with the 600Ω input Z of the LCR presents a 300Ω load to the pentode which throws away precious gain. Granted a pentode loaded with 300Ω will have really low distortion but I only get around a gain of 2 from a 7788 loaded with 300Ω.... taking the LCR Z and plate load up to 3100Ω will net 12dB more gain out of the stage bringing it up to over 40dB rather than the ~32dB it has as drawn

dave
Get Your Fix
www.hifiheroin.com
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

In the last schematic the EQ600 is not terminated with 600 ohm so load seen by input stage is higher
The another ptoblem is the overload of ths circuit; with 5 mV at 1 kHz the 30dB are too low for me

Walter
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

hello walter.... you are the one big on actual measured results in place of simulations so measure the input Z of a 600Ω LCR with and without the 600Ω termination and let me know if you want to change your above statement.

dave
Get Your Fix
www.hifiheroin.com
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

You have 300 ohm only if you have Zs 600 ohm and Zload(network) 600 ohm.
Then if you look at the graph I sent you can see the response
Zs=600 ohm
Zload= 100k, 47k, 10k

The blu at 0 dB line is Zs=600 ohm and Zload =600 ohm

This is a LCR network only


Walter
Attachments
LCR_resp_Zload_100k_47k_10K_600_Zs_600.jpg
LCR_resp_Zload_100k_47k_10K_600_Zs_600.jpg (260.73 KiB) Viewed 77257 times
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

One add.
In the Th. schematic there aren't the 600 ohm in input and 600 ohm in output of the LCR.
The EQ600P the 600 ohm are outside the LCR circut (that is inside the box).
In this case the Zout of the gain stage is 600 ohm, it is a Zin of network, but the load is the 1,5 Mohm (in ac).
And, as the graph says, 6 dB of more gain is helpful
But the flat response is not assured

Walter
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

I do not know the conditions of your test but it shows the output in the frequency domain and I asked you to measure the Z-in of the LCR with 1.5MΩ and 600Ω terminations.

dave
Get Your Fix
www.hifiheroin.com
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

Hi

the condition of my the test is described.
The Zin referred ay 1 kHz is 600 ohm
The 6 dB of loss when I move from an high load (33 or 47K, 10K the loss is a little bit lower) to 600 ohm give us the Z

Note: one error on my post, the Zload is the parallel of 1M5 and the Z of network.

Note2 = the Th. way of connection the EQ600P (without cap) with first stage in the standard Tango configuration was not expected; while with EQ-2L yes ( no resistors in parallel with cap to ground as the example shown) but the accuracy is less declared. So a little dc current flow in the inductors, I guess
But the solution is interesting.

Just for info in attach the answer of my LCR stage

Walter
Attachments
LCR_risp.jpg
LCR_risp.jpg (150.71 KiB) Viewed 77239 times
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Since you do not want to make the measurement I asked you for, I will tell you that the Z-in of a 600Ω LCR network above 1kHz is 600Ω independent of any terminating resistor. load it with 600Ω and the Z-in is 600Ω, load it with 1M5 and the Z-in is 600Ω. There is a small increase in Z below 100hz but this is trivial for this situation.

this was all in response to your comment that the pentode in the referenced phono stage will be loaded differently with a 600Ω and a 1M5Ω termination of the network.

dave
Attachments
Rsource = 600Ω   <br />Treminating R= 600Ω green, 1m5Ω Blue
Rsource = 600Ω
Treminating R= 600Ω green, 1m5Ω Blue
Screen Shot 2021-06-03 at 5.19.12 PM.png (25.02 KiB) Viewed 77233 times
Get Your Fix
www.hifiheroin.com
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

Hi

please don't write things not real.
I never said the I don't want to make a test.
And the diagram with curves is itsef a test (real) where you can find the Z so your simulation is not necessary.
I wrote also a little explanation
To be precise, I wrote the reference at 1 kHz as reference point; it is evident that the linearity of curves says the the Zload is costant with frequency

In my previuos post I also put a note about my error, maybe you haven't read carefully.
And I gave you an answer about Z.

Walter
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

In the last schematic the EQ600 is not terminated with 600 ohm so load seen by input stage is higher
this entire branch of the discussion was in response to your quoted text above.

I believe the quoted text to be incorrect which is why I asked you to simply measure the input Z of the EQ-600 with 600Ω and 1M5Ω.

If you still believe your quoted text above I will simply ask you to tell me the input Z of the network under those two conditions. To keep things simple you can choose 1kHz as a frequency.

If you think this is unimportant, then you fail to understand that this provides a load to a pentode where the gain is basically the Gm/load and for a pentode stage feeding a LCR the gain. If you would build and measure the circuit you would find the voltage at the input of the LCR the same in both conditions and the difference in your measured outputs simply has to do with loading a 600Ω LCR with 600Ω gives the ~6dB loss you show. The LCR is a T network which has the roughly the same input and output Z characteristics independent of the source and load conditions. There is a slight deviation from the Ideal "T" at low frequencies as shown in the plots above.

dave
Get Your Fix
www.hifiheroin.com
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

The text is correct.
The Tango data sheet for EQ600P show the network closed to 600 ohm.
Other configurations ( as I use, in case) aren't compliant with the Tango indication.
If you don't close the network with 600 ohm the Z is higher respect the original diagram.
This can be seen from my diagram; maybe wolud be better to connect other resistor as 5kohm and 3 kohm (just an example) to show other variations of level.
I have sent here another diagram on other thread with the test changing the Zource leaving the Zload of network at 600 ohm and the shape of curves are similar to the last ones but a little bit better.

This is my diagram, is not complicated but you can configure the tubes and the load in input for LCR stage and, in case, variyng the load
Using different tube you can get, with minor changes, different gain until 46-48 dB
The use of parlle is to make easy the reach of the nominal value of the network components
There is a matrix to configure the filaments ( in the new version of pcb is is easier)


The phot of the proto for tests

Last thing about 1 kHz; I mentioned 1 kHz, as specified, as reference because with Audio Precision I will get the reference, then I will start the sweep.
I don't know if yu are confident with test set but normally with this stuff ( and others)

Walter
Attachments
Figura 1_LCR_Fono_schema.jpg
Figura 1_LCR_Fono_schema.jpg (241.69 KiB) Viewed 77222 times
Figura 9_proto_2 (1).JPG
Figura 9_proto_2 (1).JPG (1.07 MiB) Viewed 77222 times
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

If you don't close the network with 600 ohm the Z is higher respect the original diagram.
Would you please measure this so you can realize that you are incorrect. If you consider a 600Ω LCR as a three terminal "black box" with in, out and ground, it has an 600Ω input AND output impedance. It is a constant impedance "T" network and whatever happens on the output is largely invisible to the input and whatever happens at the input is largely invisible to the output.
This can be seen from my diagram; maybe wolud be better to connect other resistor as 5kohm and 3 kohm (just an example) to show other variations of level.
The change in level you see from your test has nothing to do with the load the LCR presents to the source and everything to do with the fact it has a 600Ω output impedance. The ~6dB difference you see between a 100kΩ and a 600Ω termination is simply proof that when Rsource=Rload you lose 6dB.

Whatever value you terminate the LCR with will simply attenuate based on a 600Ω source feeding that value load. 600Ω is 6dB, 1kΩ is 4dB, 1500Ω is 3dB and 5000Ω is 1dB. in each and every one of the cases above say 100hz the load presented to the source by the input of the LCR will be 600Ω

dave
Get Your Fix
www.hifiheroin.com
docali
Posts: 58
Joined: Thu Oct 16, 2008 7:16 am

Post by docali »

Dave's comments are correct for me. The LCR has two bridged T-networks. rtfl, where l stands for literature :wink:

Walter's view may be valid for near DC conditions.
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

Maybe you haven't read my previous post where I specified my error.
Then the input of circuit connected to out of the LCR network see the 600 ohm in parallel, in case of TH. design, with 1M5.
If I put , p.e. 10k, the circuit see 600//10k.
In my case I decided to use 600 ohm as load of LCR and a variable Zsource ( in a range as plot on test lab)because the frequency answer is a little bit better respect to Zs= 600 ohm and the LCR close with high values
The difference in 0,2 dB is not a problem but the goal was to reach the best possible.
And it is evident that the accuracy on values of the inductors, resistors and caps make a difference.
For this reason I have planned the possibility to put in parallel two or more components to reach the right value ( or close to it)

Walter
westernelectric300b
Posts: 28
Joined: Fri Apr 30, 2021 1:29 am

Re: Step-up after LCR

Post by westernelectric300b »

Hi Dave Slagle,

So to cut to the chase, are you suggesting that the terminating Z at the output of the 600 ohm LCR network is essentially irrelevant to the accuracy of the RIAA curve provided by the network, and a terminating value of (example) 600 ohms only serves to attenuate the output of the network?

I had planned on terminating the output of the LCR network in my design with a load Z of 600 ohms, thinking this was essential to preserve the accuracy of the network, but if I understand your logic, I can terminate the same network in say 5000 ohms, and in so doing, I'll retain the RIAA accuracy, but the overall gain of the system will be improved by 5 dB.

Am I understanding you correctly?

Thanks & Regards,

WE300B
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Re: Step-up after LCR

Post by dave slagle »

The LCR network either needs to be driven by its characteristic impedance or loaded at its characteristic impedance. In either case you have a situation where Rsource=Rload which causes a 6dB loss of amplitude so there is really no difference in gain either way you go about it.

This all started when Thorstens pentode version was brought up where he loads the pentode with 600Ω to give a 600Ω source to allow for the output to remain unterminated so he could use a tiny silver mica coupling cap. My problem with this topology is the 600Ω load resistor appears in parallel with the 600Ω input Z of the LCR to load the pentode with 300Ω. This means you get a 6dB loss from the Rsource=Rload and halving the load to a pentode also reduces gain by 6dB for a 12dB total loss. Moving to a 1200Ω load and a 1200Ω LCR would gain 6dB back.... 2400Ω on each gets another 6dB.

dave
Get Your Fix
www.hifiheroin.com
westernelectric300b
Posts: 28
Joined: Fri Apr 30, 2021 1:29 am

Re: Step-up after LCR

Post by westernelectric300b »

Dave Slagle,

That makes perfect sense.

In my case, the output of the 600 ohm LCR network is driving a transformer (Sowter type 9063) with a 600 ohm primary, and the secondary Z of the transformer is 10K ohms. Terminating the secondary of this transformer in a 10K load resistance (which is actually the value of the grid leak resistor of the 2nd gain stage) will reflect a 600 ohm load to the output of the LCR network, so I should be fine. As such, I am unconcerned about the source Z driving the LCR network; I am only concerned about the load Z that the network is looking into.

Do you agree with my logic?

Thanks & Regards,

WE300B
westernelectric300b
Posts: 28
Joined: Fri Apr 30, 2021 1:29 am

Re: Step-up after LCR

Post by westernelectric300b »

Hello Pecci!

Please excuse my very late reply; today was the first day that I have visited this site since the very early part of June, and I completely missed your post!

Thank you for posting the information concerning the fixed biasing arrangement. That is very interesting. I think for these very high-u, high Gm tubes, fixed bias is a very promising approach. It eliminates the need for cathode bypass capacitors, which I never use in any of my audio designs, and it allows for precise biasing to obtain the required Ip of the tube. The alternative, of course, is either to use LED biasing, or battery biasing. Personally, I never liked battery bias, and I'm not all that enthusiastic about LED biasing, either, but I think that for the first iteration of my LCR phono preamp, I will be using the LED bias for the D3As. With an Ip of 20 Ma, the LED bias should work quite well.

I think that any IR (infrared) LED will work well with the D3A, at the desired Ip/Ep that I will be using. An IR LED typically has a forward voltage drop on the order of 1.2 volts, and I will probably have to select-in-test a number of different IR LEDs in order to find those that provide the precise Ek we need to properly bias the D3As. Fortunately, these LEDs are very inexpensive, so it makes sense to buy a bunch of them through eBay, etc., and select those that provide the ideal operating point for the D3As. The forward voltage drop across the LED increases somewhat with increasing current, so the IR LEDs may be perfect for this application.

The 4 month lead time from Sowter for the 9063 step-up transformers is very long. I ordered mine from Sowter around May 15th, and they quoted delivery of around early August. I hope the delivery does not stretch much beyond that, as I would like to move forward with building my LCR phono preamp.

The metalwork for my LCR phono preamp is completed, and I may begin assembling the unit within the next few weeks. Like all of my audio creations, the PSU is built on a separate chassis, and it will use (4) TV damper diodes in a bridge rectifier configuration. Lundahl is closed for the month of July for vacation, and they and Sowter will be the suppliers that will be holding up the completion of my project.

All the Best,

WE300B
Last edited by westernelectric300b on Tue Jun 29, 2021 9:20 pm, edited 3 times in total.
westernelectric300b
Posts: 28
Joined: Fri Apr 30, 2021 1:29 am

Re:

Post by westernelectric300b »

pecci wrote: Tue Jun 01, 2021 6:19 pm Hello WE300B,
I attach schematic of Thorsten LCR. You can find their full description of that preamp on Jogis Roehrenbunde site, but in German only. Please note SUT on input side missing in schematic and Thorsten use S&B TX103.
Additionally, I add one schematic from Neumann, just for illustration of that biasing concept.
I don’t know what is the correct name of that bias but that direction looks very promising for me and I will be deeply grateful for any help on that road.
Key point here is to biasing tube (especially high Gm types like D3a) on elastic and floating way by mutual interaction of voltage dividers from plate and ht supply to achieve better musicality with more relaxed sound, and partly avoid painful tube matching.
Your OP for D3a is close to my, maybe I prefer slightly lower Ib and Ub for about 10%.
And yes, lead time from transformer producer become longer and longer. I was waiting 9063 for almost 4 months. In comparison to them Dave is clear champion. Thanks Dave.
What is your suggestion regarding LED type?
Best regards,
Hello Pecci!

Please excuse my very late reply; today was the first day that I have visited this site since the very early part of June, and I completely missed your post!

Thank you for posting the information concerning the fixed biasing arrangement. That is very interesting. I think for these very high-u, high Gm tubes, fixed bias is a very promising approach. It eliminates the need for cathode bypass capacitors, which I never use in any of my audio designs, and it allows for precise biasing to obtain the required Ip of the tube. The alternative, of course, is either to use LED biasing, or battery biasing. Personally, I never liked battery bias, and I'm not all that enthusiastic about LED biasing, either, but I think that for the first iteration of my LCR phono preamp, I will be using the LED bias for the D3As. With an Ip of 20 Ma, the LED bias should work quite well.

I think that any IR (infrared) LED will work well for with the D3A, at the desired Ip/Ep that I will be using. An IR LED typically has a forward voltage drop on the order of 1.2 volts, and I will probably have to select-in-test a number of different IR LEDs in order to find those that provide the precise Ek we need to properly bias the D3As. Fortunately, these LEDs are very inexpensive, so it makes sense to buy a bunch of them through eBay, etc., and select those that provide the ideal operating point for the D3As. The forward voltage drop across the LED increases somewhat with increasing current, so the IR LEDs may be perfect for this application.

The 4 month lead time from Sowter for the 9063 step-up transformers is very long. I ordered mine from Sowter around May 15th, and they quoted delivery of around early August. I hope the delivery does not stretch much beyond that, as I would like to move forward with building my LCR phono preamp.

The metalwork for my LCR phono preamp is completed, and I may begin assembling the unit within the next few weeks. Like all of my audio creations, the PSU is built on a separate chassis, and it will use (4) TV damper diodes in a bridge rectifier configuration. Lundahl is closed for the month of July for vacation, and they and Sowter will be the suppliers that will be holding up the completion of my project.

All the Best,

WE300B
Last edited by westernelectric300b on Tue Jun 29, 2021 9:20 pm, edited 1 time in total.
Post Reply