Step-up after LCR

the road not taken.
Stegie
Posts: 10
Joined: Tue Nov 10, 2020 11:05 am

Step-up after LCR

Post by Stegie »

I have been following the forum for quite some time now and have read numerous topics with great interest. Thank you and hello to you all!

I am at the moment working on a phono design and got stuck in my ambitions. Perhaps they are too high, or perhaps I am overlooking something.

My goal is a 2-stage, high gm, 600 Ω-LCR phono with 40-ish dB gain, low Z-output and without any (local) feedback.

Achieving this with a pentode in the first stage is the easy way out. Same goes for 3-stages. But I aim for a triode in the first stage, 2 stages total, and got inspired by a couple of examples that use a step-up. However, I am still puzzled how they made this work, or if they did accurately at all.

My design so-far:
- 1-st stage (D3a or EC8010)
- 3:1 / 4:1 step-down to load the LCR
- 1:4 step-up
- 2-nd stage similar to first
- 2:1 step-down for low Z-out

The step-up after the LCR will have a low primary impedance. I expect to achieve 15-25H max. Consequently the primary impedance will fall with decreasing frequency, and as a result:
- the R-load that the LCR ‘sees’ will drop to 80%. (my perception is that the RIAA will no longer be accurate at low frequencies)
- the voltage divider between LCR and step-up becomes 7% smaller. This results in 0.6 dB loss of bass.

Am I overlooking something? Or am I being overly theoretical?

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

Post by dave slagle »

I have a friend that insists that using a 10K:600 >>> 600Ω Daven T ≥≥≥ 600:10K was the best resistive based volume control he has heard. This concept isn't far off from what you suggest.

the thing that helps you a bit in this is assuming you set the impedance of the LCR on the secondary then the output Z of the network will be 300Ω making the Step up an easier device to build.

dave
Get Your Fix
www.hifiheroin.com
Stegie
Posts: 10
Joined: Tue Nov 10, 2020 11:05 am

Post by Stegie »

I had already foreseen that the output Z of the network would be 300Ω, and that that would increase the low-frequency response of the step-up.

However, the primairy impedence will fall with frequency and will shunt part of the reflected load. Surely your friend lost some gain at low frequency?

The same mechanism will effect the stability/load of the LCR. Could you help me with your thoughts on that?

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

Post by dave slagle »

Losing gain at low frequency is just part of the deal with the devil that you accept when you use transformers. I would simply spin the whole thing and call it a rumble filter :-)

attached are the plots of the pultec 600R LCR loaded with 20, 40 and 100hy. I am not bothered by -1dB @ 7hz... particularly in a phono front end.

dave
Attachments
Screen Shot 2020-11-10 at 5.18.55 PM.png
Screen Shot 2020-11-10 at 5.18.55 PM.png (60.51 KiB) Viewed 94659 times
Get Your Fix
www.hifiheroin.com
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

Hi
I have read this thread plus others about LCR
Now I am working with a friend of mine ( owner of 105 Moving coil cartridges) around LCR stage with all original Tango stuff. (different models included 999)
The use of step up after riaa stage was suggested to him by me in case of low gain.
Now we are in testing the phono also under musical aspect ( my friend has a heavy experience on live concerts).
At the moment the results are very good.

One annotation.
I have seen a lot of threads about LCR but the real test are missing.
Only simulation; in case help me to find some real results done from someone.
I have done dozen of test in different configuration where the LCR (original EQ 2L + external components + different IT trafos) was involved ( also with Shishido circuit).
I have two Audio Precision (Sys One + Sys Two with riaa test wrote by BBC in UK!!!) , three generatos, five tube tester plus much other equipments.
After the end of the test I can provide, if my friend agree with me, some real result.

Ciao

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

Post by dave slagle »

I have found that builds based on spice simulations tend to closely follow the simulation if all of the models used are reasonably accurate. In almost all cases the errors that do show up come from inaccuracies or flaws in the models. One a working physical model is built based on a simulation the simulation model parameters can be altered to match the real world and surprisingly consistent results can be realized.

With RIAA based circuits I always insist people have a working spice simulation and a swept real world example for comparison. Ultimately the actual sweep is correct but the spice simulation will tell you what to change to move in the direction you need to go.

At no point am I suggesting that simply copying a spice simulation will net perfect results. Spice will give a great starting point and then becomes an incredible tool for quickly effecting changes in circuit.

Here is a sim of an inductive based RIAA


Image

Here is a measured response of that circuit. The differences above 10Khz between the two comes down to the simplified model of capacitance in the inductors compared to the complex real world reality. The difference in the low frequency response can be attributed to the real world damping of a resonance that isn't accounted for in the spice models.


Image

and here is the simmed response with the inductors swept through a number of values and if the real world model is swept through those values the pattern will be the same.


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

Post by walge »

Hi

thx for answer.
I am looking for a real test.
And I suspect that nothing of them are available for LCR circuits.
I haven't nothing against simulation.
But it is easy to produce circuits and diagram only by simulation.
For this heavy project I spent about 50 hrs only on test labs with hundreds of measurement
Of course, modifying, where necessary, the circuit.
Heavy job but with a lot of satisfaction.

each channel, 5 Tango original stuff + 2 x Ec8020 ( and some other)


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

Post by dave slagle »

I am not sure how much more of a real test you need than a frequency sweep of an actual circuit that has been built that I showed along with the simulations of the same circuit.

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

Post by walge »

Hi

there are some additional real test.
The Thd vs frequency, max acceptance vs frequency.
FFT at different frequency
ratio s/n

In addititon the real test on the L involved in LCR.

Walter
Stegie
Posts: 10
Joined: Tue Nov 10, 2020 11:05 am

Phase shift

Post by Stegie »

Hi Dave,

I hadn't touched Spice until you set me on track in november and I am having a great deal of fun and insight since. Thank you for that!

I agree that the rolling-of at low frequencies is not that big of a deal, but how about the associated phase shift? Usually phase-shift is regarded as a negative and I reckon in audio it is the same, although it is hard to find any information on the matter.
When using (interstage) transformers throughout a total system, the phase-shift ads up to rather large numbers. Wouldn't that mess up the transient response? Or are there ways in transformer design to deal with this?
Attachments
My intitial goal: a LCR interstage phono
My intitial goal: a LCR interstage phono
EST - Phono LCR0600 - D3a - B.jpg (280.3 KiB) Viewed 94484 times
Gain 39 dB<br />30 degrees phase shift<br />0.5 dB roll-of at low frequency<br />Rout 260 ohm
Gain 39 dB
30 degrees phase shift
0.5 dB roll-of at low frequency
Rout 260 ohm
EST - Phono LCR0600 - D3a - C.jpg (222.45 KiB) Viewed 94484 times
EST - Phono LCR0600 - D3a - A.asc
(15.88 KiB) Downloaded 188 times
docali
Posts: 58
Joined: Thu Oct 16, 2008 7:16 am

Post by docali »

Stegie
Posts: 10
Joined: Tue Nov 10, 2020 11:05 am

Post by Stegie »

This is quite a coincidence. It was your paper that started me off in this direction in the first place! Thank you for your study, and thank you for bringing it under my attention once more!

The impact of misalignment between input and output resistance for the LCR is now very clear to me. To the contrary of certain remarks and opinions, it does effect the linearity of the LCR and it has a more-or less hidden effect on the linearity of the step-up after the LCR: the misalignment leads to additional low frequency roll-of and associated phase-shift.

I am happy with the result: roll-of at low frequency has really improved, but more importantly, phase shift has improved from 30 to 20 degrees.

As this is my first design with interstages, the phase shift is still rather large as what I am used to. Perhaps I am overly cautious, but when using (interstage) transformers throughout a total system, the phase-shift ads up to rather large numbers. Wouldn't that mess up the transient response? It is not that easy to get a grip on this, and I would really appreciate some friendly advice!

Erik
Attachments
EST - Phono LCR1200 - D3a - B.png
EST - Phono LCR1200 - D3a - B.png (224.21 KiB) Viewed 94466 times
EST - Phono LCR1200 - D3a - C.png
EST - Phono LCR1200 - D3a - C.png (225.67 KiB) Viewed 94466 times
EST - Phono LCR1200 - D3a - A.asc
(24.85 KiB) Downloaded 181 times
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

To Docali

please send me in private your email
I mentioned your paper in my aritcle on Lcr I am writing
In next steps I will have only real lab mesurament on LCR stage in different topologies (already tested quickly)

Walter
docali
Posts: 58
Joined: Thu Oct 16, 2008 7:16 am

Post by docali »

something secret or better discuss it here?
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

no secret.
:)



Bye

Walter
pecci
Posts: 4
Joined: Sun Jan 03, 2021 6:57 pm

Post by pecci »

Hi,
Please, please continue here.
Its evident that step up after LCR must be of significant size and inductivity and not some tiny types.
But still remain to explore influence of:
secondary termination load on step up performance and LCR accuracy, and
applying dc through step up secondary for biasing next D3a tube and same thing for MC step up for first stage.
I just prepare to start assembling one LCR600 with same topology and all this still remain as open questions for me.
Br
Pedja
walge
Posts: 36
Joined: Tue Nov 24, 2020 8:42 pm
Location: roma

Post by walge »

Hi have some test lab with a step up after Lcr then connected to next gain stage.
Good results but not perfect.

I am trying other similar solution with different stuff

Walter
Stegie
Posts: 10
Joined: Tue Nov 10, 2020 11:05 am

Post by Stegie »

Mr. Docali's white-paper set me of on a certain track, by using a 1200 ohm LCR and a 1200 ohm termination resistor. These in parallel, offer the step-up with the proper 600 ohm load, according to its design specs.

Sounded like a sensible proposition, but something kept stirring and didn't feel right. Also I didn't like the strange frequency response that I got.

Mr. Docali: are you really sure about your white paper?
1. how about the reflected impedance, that will be seen by the LCR in parallel with the termination resistor? The step-up's secondary should be terminated with a very high resistor to present a reasonable reflected impedance at the primary side, but at the cost of high frequency loss.
2. and on a side note: presenting 600 ohm to the stepup is according to general specs, but when was it ever a negative to present a lower impedance to a transformer?

My solution only requires two actions:
1. ditch the termination resistor at the output of the LCR
2. use a 600 ohm LCR and have that loaded at it's output by the step-up transformer. This way, both LCR and step-up see a 600 ohm load (depending on step-up ratio and primary and secondary DCR resistances, the right termination resistor can be calculated to load the step-up (on the secondary) and reflect 600 ohm to the primary.

As a result:
1. the frequency response looks beautiful, without any trace of resonances
2. @30-20kHz: roll-of at low and high frequency's is less than 0.1 dB
3. @30-20kHz: phase response 14 / -8 degrees
Attachments
Response LCR600 30-20k.png
Response LCR600 30-20k.png (214.31 KiB) Viewed 94329 times
Response LCR600 30-20k.png
Response LCR600 30-20k.png (214.31 KiB) Viewed 94329 times
Circuit.jpg
Circuit.jpg (287.09 KiB) Viewed 94329 times
Stegie
Posts: 10
Joined: Tue Nov 10, 2020 11:05 am

Post by Stegie »

I also compared the 600, 1200 and 2200 ohm LCR in my circuit, al things else being equal:

LCR-600:
- gain: 40.2 dB
- 30 Hz: 0.07 dB roll-of
- 20 kHz: 0.084 dB roll-of
- phase shift 14 / -8 degrees

LCR-1200
- gain: 44 dB
- 30 Hz: 0.29 dB roll-of
- 20 kHz: 0.32 dB roll-of
- phase shift 18 / -16 degrees

LCR-2200
- gain: 45.3 dB
- 30 Hz: 1.51 dB roll-of
- 20 kHz: 1.02 dB roll-of
- phase shift 33 / -27 degrees

The higher the LCR-resistance, the higher the required a termination on the secondary of the step-up, and subsequently at the cost of high frequency response.

Haven't found the reason (yet) for the low-frequency variation. Perhaps anyone is willing to give his thoughts?
Attachments
Response LCR2200 30-20k.png
Response LCR2200 30-20k.png (153.15 KiB) Viewed 94328 times
Response LCR1200 30-20k.png
Response LCR1200 30-20k.png (161.58 KiB) Viewed 94328 times
Response LCR600 30-20k.png
Response LCR600 30-20k.png (150.18 KiB) Viewed 94328 times
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

images are too big... if you scale them down to fit the nasty scroll of death will not happen :-)

for the loss of bass, did you increase the inductance of the transformers following the LCR as you went up in impedance? For s fixed inductance, I would expect the 600Ω version he go two nearly octaves lower than the 2200Ω version.

dave
Get Your Fix
www.hifiheroin.com
Stegie
Posts: 10
Joined: Tue Nov 10, 2020 11:05 am

Post by Stegie »

Hi dave,

I noticed that my images were rather big. Here comes another try :-)

I didn't change anything, other then the inductance of the LCR and the termination resistor on the secondary side of the step-up in order to create the matching load. The loss of base, on this level, is not an issue. I just wondering what mechanism could cause it.

I am pretty contempt with the circuit, if it wasn't for:
a> Miller capacitance is rather big. Not important for MC-cartridges, but it might rule some MM-cartridges out. Could be acceptable
b> BUT there is not much signal left at the end of the LCR module:
- 5 mV goes into the 1st. stage / SNR 82 dB is possible (2 ohm cartridge, 1:10 SUT)
- 7 mV comes out of the LCR-module, and there are some pretty big resistors in contact with that tiny signal. I calculated a SNR = 64,5 dB

Makes one wonder how the 'professionals' design their all triode 2 stage LCR phonos. The same tube, wired as pentode, might solve both issues.

What are your thoughts on that?

Erik
Attachments
Schermafbeelding 2021-02-09 om 21.24.15.png
Schermafbeelding 2021-02-09 om 21.24.15.png (119.52 KiB) Viewed 94312 times
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

you need to increase the primary inductance of the following transformer to account for the increasing LCR impedance. a 1200Ω LCR requires a following transformer with double the inductance for the same low end behavior. A 2200Ω LCR will require just under 4X the inductance. Keeping the inductance constant will result in an octave less of bottom end for each doubling of the LCR Z.

as an additional detail to consider, the loading of the secondary of a transformer to reflect back the required terminating Z for the LCR is not IMO a wise choice.

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

Post by docali »

[quote="Stegie"]
Mr. Docali: are you really sure about your white paper?
1. how about the reflected impedance, that will be seen by the LCR in parallel with the termination resistor? The step-up's secondary should be terminated with a very high resistor to present a reasonable reflected impedance at the primary side, but at the cost of high frequency loss.
2. and on a side note: presenting 600 ohm to the stepup is according to general specs, but when was it ever a negative to present a lower impedance to a transformer?
[/quote]

Hi,

yes, i am sure. You simply have to read it carefully. Obviously, you neither have understood my intention to use an LCR1k2 nor how to apply the circuit with a following step-up transformer.

Every transformer has a self resonance and I provided measurements in my paper for the Sowter in dependence of the driving impedance. You do not need to follow this road to damp the internal resonance. You do not need to follow the road to add a step-up behind the LCR. But if you do then do it right.

The LCR need to be terminated almost correctly before the step-up. To remove the resistor after the LCR is your fault as this greatly increases the driving impedance. If I remember well I used extensively the notation driving impedance in my paper which is important to damp the step-up's resonance.

The low resistor after the step-up is your fault. It is considerably below the rated impedance of the secondary and will of course tame the HF response. You can leave the secondary open if the driving impedance is appropriate.

And as Dave mentioned, you need to make sure that the step-up is able to work with a driving impedance of 600 ohms! For the Sowter I checked this carefully and you also already see this by inspecting the datasheet.

Think about the load line for the D3a by using the LL1692a as 4:1.75. I would use it as 3.5:1.

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

Step-Up After LCR

Post by westernelectric300b »

With regard to the two-stage LCR phono preamp schematic posted here with the two D3a tubes; please advise what symbols V1 and V2 are. Are these 1.3 volt LEDs, diodes, batteries, etc.?

Also, what is symbol E1? Is this the RCA phono input connector, or the MC-SUT, etc.?

And lastly, what is the gain of this unit at 1 Khz?

This appears to be a very interesting design, and I am interested in building one of these preamps for my own use.

Thanks very much!
Last edited by westernelectric300b on Fri May 07, 2021 5:01 am, edited 1 time in total.
westernelectric300b
Posts: 28
Joined: Fri Apr 30, 2021 1:29 am

LCR Phono Preamp

Post by westernelectric300b »

Has anyone built this unit to the schematic diagram that was posted here?

If so, I'd be curious as to the actual measured data, and the sonic performance.
westernelectric300b
Posts: 28
Joined: Fri Apr 30, 2021 1:29 am

LCR Phono Preamp

Post by westernelectric300b »

I think I answered my own question about the purpose of schematic symbol E1; it is obviously the MC-SUT and it adds 20 dB of voltage gain.

But schematic symbols V1 and V2 are still intriguing. Are these LEDs for cathode biasing of the tubes? Perhaps schematic symbol V is a European designation for a diode, etc. Here in the U.S., we of course use the designation D for diodes, including LEDs.
westernelectric300b
Posts: 28
Joined: Fri Apr 30, 2021 1:29 am

Step-Up After LCR

Post by westernelectric300b »

Further to my earlier postings;

I plan on building this unit, but with a few possible changes to the schematic that was posted here.

I assume that symbols V1 and V2 are LEDs for biasing of the tubes. I may try this approach, but I also plan on trying to use a cathode resistor with the so-called Ultrapath approach, to eliminate any need for a cathode bypass capacitor. I have used Ultrapath to great success in the past with my 2-stage DHT power amplifiers, but never with tubes of high u such as the D3A or WE-417A. The issue here of course is any residual power supply ripple is amplified by the u of the tubes, and reducing the PSU ripple to a level where this is not an issue in a high-gain phono preamp may be a problem. I was never a big enthusiast of LEDs for biasing a tube, but they are a workable solution in a phono preamp, so I'll give them a listen.

I also plan on using a pair of the S & B LCR EQ modules that I purchased from them around 12 years ago, when S & B was still supplying them. They are 600 ohm devices, and from what I recall, are electrically identical to the Tango EQP-600 units.

I'll use the D3A for the first gain stage, and the second gain stage will use a WE-417A. The 5 dB reduction in overall gain with the WE-417A will be compensated for by the MC-SUT I'll be using, which provides around 23 dB of voltage gain after the predicted insertion loss of the transformer. And I have a lot of WE-417As here.....

And has anyone decided which is a better choice for the step-up transformer after the LCR network? A Sowter 9062, with a Z ratio of 600/40K, or the Sowter 9063, with a Z ratio of 600/10K?
Last edited by westernelectric300b on Fri May 07, 2021 5:01 am, edited 1 time in total.
westernelectric300b
Posts: 28
Joined: Fri Apr 30, 2021 1:29 am

Step-Up After LCR

Post by westernelectric300b »

Gee, I'm really surprised that after a week, I have not received any replies to my inquiries here!
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Hey,

I think you have answered all of your own questions and it is not my design so it is not my place to comment.

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

Step-Up After LCR

Post by westernelectric300b »

Hi Dave,

Thanks very much for the reply! I was beginning to wonder if anyone had seen any of the posts I had made on this topic.

Perhaps you can answer a question ....reference designations V1 and V2 on the schematic are a little confusing to me. Here in the U.S., these designations of course refer to vacuum tubes, but on the schematic, I believe they refer to either batteries or LEDs for biasing of the D3As. My guess is that they are LEDs, but I cannot be sure. Do you know what they are?

And have you had any experience with LED bias of these tubes when they are used in a phono preamp? It would seem to me that LEDs would be ideal in this application, due to the cathode current drawn through them by the D3As (18 to 20 Ma), and the obvious elimination of cathode bypass capacitors.

I'd like to try an unbypassed resistor on the cathodes of the D3As, but the Rp of the tubes rises to the point (around 8K ohms or so) where getting transformers with a high enough primary impedance becomes something of a problem, along with the loss of gain.

Ultrapath is probably out of the question, as the mu of the D3As is so high that any PSU ripple will be amplified by the tube by mu+1, and hum modulation will be an issue. A high value electrolytic from cathode to ground with the Ultrapath would take care of that, but we're back to a poor sounding electrolytic in the signal path. Hence the possible desirability of the LED bias with the D3As......
Post Reply