I have my SCA-35 for couple years. I recapped it right after I got it. Even though, I am not happy about it.
There are couple of issues.
1. It is very picky about 7199 tubes. I tried couple 7199s. Sylvania ones seem to be better than RCA ones. Anyway, my goal is to mod it to use more common tubes like 6BL8. Look at the schematic, R29 is 4.7mega Ohm. Not all 7199s or the substitutes are happy about the "Leak Bias" setup.
2. I don't have a turn table yet, but I still want to keep the phono stage and tone control intact. In case I want to try them in the future. Thus converting it into ST35 is not an option.
PS: Let's focus on the amp section and please leave the phono and the power supply mod out of this discussion.
Edit: Final Version #40
There are couple of issues.
1. It is very picky about 7199 tubes. I tried couple 7199s. Sylvania ones seem to be better than RCA ones. Anyway, my goal is to mod it to use more common tubes like 6BL8. Look at the schematic, R29 is 4.7mega Ohm. Not all 7199s or the substitutes are happy about the "Leak Bias" setup.
2. I don't have a turn table yet, but I still want to keep the phono stage and tone control intact. In case I want to try them in the future. Thus converting it into ST35 is not an option.
PS: Let's focus on the amp section and please leave the phono and the power supply mod out of this discussion.
Edit: Final Version #40
Last edited:
Go over to Dave Gillespie’s store and get the PC-10A PCBs. It ticks off all your check marks. http://www.tronola.com/html/daves_store.html
And while you are there be sure to study the benefits of doing the power supply with EFB (tm). http://www.tronola.com/A_New_Look_At_An_Old_Friend.pdf
Installing the power supply and the PC-10A modules at the same time saves some work. I doubt that you will get a better SCA-35 than that, except see below.
Dave recently published some further modifications and improved the SCA-35 to a new level of performance. You can read about those here:
https://audiokarma.org/forums/index...g-out-the-very-best-of-dynacos-sca-35.999693/
And while you are there be sure to study the benefits of doing the power supply with EFB (tm). http://www.tronola.com/A_New_Look_At_An_Old_Friend.pdf
Installing the power supply and the PC-10A modules at the same time saves some work. I doubt that you will get a better SCA-35 than that, except see below.
Dave recently published some further modifications and improved the SCA-35 to a new level of performance. You can read about those here:
https://audiokarma.org/forums/index...g-out-the-very-best-of-dynacos-sca-35.999693/
Last edited:
https://audiokarma.org/forums/index...g-out-the-very-best-of-dynacos-sca-35.999693/
ETA: Francois beat me to it. ;-)
ETA: Francois beat me to it. ;-)
The discussion from Audio Karma is very informative, although I don't like the "Karma" website with aggressive popup advertising. Thanks for sharing.
From the original schematic, the voltage at the plate of the 7199 pentode is marked only 55v. Is it too low? I thought it should be 100v or so for healthy operation.
The D. Gillespie Designs PC-10A is a very good base line to start with. The 4.7M input resistor may still cause issue for non-7199 tubes. I would short the input capacitor C17 shown as below. The thing in front of C17 is just a balance pot, which has lower impedance to the ground. It provides more reliable bias for the 7199 or the substitute tubes. Leave the 4.7M there in case the pot open circuit occasionally. Also, an extra cap before the tone stack maybe necessary to block any possible DC offset.
An extra resistor may be needed, shown as below. Without the resistor, the plate voltage of the pentode may be too low. The extra resistor could be anywhere 330K to 3.3M. I would aim the plate voltage around 90v ~ 100v. Thus the triode section has enough head room.
From the original schematic, the voltage at the plate of the 7199 pentode is marked only 55v. Is it too low? I thought it should be 100v or so for healthy operation.
The D. Gillespie Designs PC-10A is a very good base line to start with. The 4.7M input resistor may still cause issue for non-7199 tubes. I would short the input capacitor C17 shown as below. The thing in front of C17 is just a balance pot, which has lower impedance to the ground. It provides more reliable bias for the 7199 or the substitute tubes. Leave the 4.7M there in case the pot open circuit occasionally. Also, an extra cap before the tone stack maybe necessary to block any possible DC offset.
An extra resistor may be needed, shown as below. Without the resistor, the plate voltage of the pentode may be too low. The extra resistor could be anywhere 330K to 3.3M. I would aim the plate voltage around 90v ~ 100v. Thus the triode section has enough head room.
Last edited:
You have to create an account for the ads to go away. But anyway, I can't speak to your questions except to say that Dave knows exactly what he's doing. Every mod I've followed from him has produced excellent results. His thread about restoring a battered and incomplete ARC D115 of mysterious provenance (turns out to have been a prototype) is really amazing.
Here is some investigation regarding the optimal bias point of variant pentode-triode tubes.
https://www.angelfire.com/electronic/funwithtubes/pentode-triode.html
That's why I think 55v plate voltage of the pentode section from SCA-35 manual is too low.
https://www.angelfire.com/electronic/funwithtubes/pentode-triode.html
That's why I think 55v plate voltage of the pentode section from SCA-35 manual is too low.
As you wish, but I reiterate that Dave gets excellent results from his boards and modifications. If he had found a reason to change it, he would have.
I fully respect what Dave did. I probably would like to buy the PC-10A board. However, let's not stop at what he had.
Here, I am trying to address the bias variant of using alternative tubes.
Why biasing through 4.7M resistor is a bad idea? See below.
https://tubes.njunis.net/?p=256&lang=en
Here, I am trying to address the bias variant of using alternative tubes.
Why biasing through 4.7M resistor is a bad idea? See below.
https://tubes.njunis.net/?p=256&lang=en
This paper referes to driving a 30w tube ( EL34 or similar) which needs larger swing. EL84 does not need this swing.
But anyway, make your mods and publish distorsion measurment. Measurment is knowing !
I looked but did not see what your objection is. Please explain what problem you see with the 4.7M resistor. What do you mean by saying “…bias through 4.7M resistor..” The pentode is not “grid leak biased”; it is biased by R31 and R29 as grid leak resistor is referencing ground.
As @petertub suggested, go ahead and make your modifications. It will be a pleasant surprise if you could further improve Dave Gillespie’s solution.
Last edited:
As said in OP, the SCA-35 is very picky about 7199 tubes. If the wrong 7199 is used, from time to time, the sound would be cut out and producing rustling noise. I was thinking there must be something wrong with the tubes, but these 7199s work perfectly fine on my Dynaco ST70. If you compare SCA35 with ST70, the 4.7M Ohm bias resistor just looks suspicious. Not mentioning swap out the 7199 for something like 6BL8 or 6GH8. They all have different bias in the circuit.
The Dave’s PC-10A mod is not hard to implement with the original board. Basicly, it just adds a RC network to the plate of the pentode section comparing to the original one. I believe I tried at some point. They didn’t resolve the bias point issue.
Does anyone have 1st hand voltage measurements of the 7199 tube? What’s the plate voltage supposed to be?
The purpose of PC-10A is to replace the now scarce and expensive (for good ones) 7199 with readily available 6GH8a or 6U8 (or ECF82). NOT 6BL8. From personal experience I know some of the NOS 6GH8as that were manufactured late in the heyday of tube manufacturing are in fact rebranded 6BL8 and they do not work right in the PC-10A. It is however optimized for genuine 6GH8a and 6U8 tubes, and they have worked flawlessly for me.
I did not experience the issues you mentioned when I used NOS 7199 in my SCA-35. Are you sure you have genuine, tested 7199 tubes? There were some problems with the Russian made 7199s sold in the late 1990, but I suspected they were not genuine 7199 and I never used them in a SCA-35. But that was before I replaced the old Dynaco PC-10 PCB with Dave’s PC-10A. I’m not understanding why you keep going back to 7199 in the SCA-35, unless you have a good supply at hand.
I did not experience the issues you mentioned when I used NOS 7199 in my SCA-35. Are you sure you have genuine, tested 7199 tubes? There were some problems with the Russian made 7199s sold in the late 1990, but I suspected they were not genuine 7199 and I never used them in a SCA-35. But that was before I replaced the old Dynaco PC-10 PCB with Dave’s PC-10A. I’m not understanding why you keep going back to 7199 in the SCA-35, unless you have a good supply at hand.
Last edited:
The Dynaco OPTs are decent. The power transformer is too small. I have 2 of each to build due mono. I am throwing out everything else but the EL84 output tubes. We built a prototype (based on HH Scott iron), and got fantastic performnce sonically. It was Class A triode so we gave up some power (3.2w at clipping) but it was compettive with my SIT-3 (so VERY good). The circuit wa based on Eli’s El Cheapo. With the dual mono ihave ambitions of a microcontoller controling things ike bias, Triode, UL, Pentode operations so that power can be traded for fidelity.
Basically an ECC81 LTP with a CCS supply driving the EL84s. The front end transfer curve tends to balance out the transfer curves of the EL84 to get more linear output. And we used all poly caps. Makes a big difference in the PSU.
dave
Basically an ECC81 LTP with a CCS supply driving the EL84s. The front end transfer curve tends to balance out the transfer curves of the EL84 to get more linear output. And we used all poly caps. Makes a big difference in the PSU.
dave
I suspect the old carbon composite resistors are failing. I did replace all the resistors on the power amp boards, except the 4.7M resistors, because I didn’t have this valve in hand. Although they measure fine when cold, they may open circuit when hot. That might explain why it ran more stable after I shorted the input coupling capacitor C17.
Last edited:
What undesirable behavior did you experience before bypassing C17? Why do you not suspect C17 as the source of the problem? Have you actually checked C17?
I don’t see why the 4.7 M resistor would cause a problem, even if it drifted some, unless it is so high that it can’t function as a grid leak. C17 and R29 form a high pass filter, and changing the component values will change the cutoff point, but not affect the proper behavior, as long as the component are intact.
I would simply get Dave’s boards (including the EFB/power supply that really makes a difference) and be done with it, but I understand you have a vested interest in salvaging the originals that you have already partially repopulated. Good luck.
Last edited:
It was not stable occasionally when hot. When it happens, the sound will be cut out and producing white noise. I suspect it is the bias issue caused by the circuit design or failing component. I also added 300K resistor shown in the post #4. The plate voltage of the pentode is too low without it. Also I added the RC network as Dave’s board in front of the phase splitter. I figured it would help to stabilize the positive feedback at high frequency.
With these symptoms i'd start with replacing the 4.7M resistor with a metalfilm ( and nullifying other changes)
Wait! You have actually installed that resistor between the HV supply and the cathode of the 7199 pentode? How much did it change the bias and performance characteristics? How much is your screen voltage now compared to the specified 32 volts. Why do you think that is too low?
The RCA datasheet has a “characteristic class A working point” with a screen voltage of 50 V. I am sure David Hafler and the Dynaco carefully evaluated their operating point and chose one that worked best. Don’t forget that a change in screen voltage dramatically changes the pentode parameters, and with direct coupling between the pentode and triode sections it could upset the whole operation.
Unless you are true expert I suggest that you undo all the “unauthorised improvements” and rebuild your PC-10 exactly as designed with good parts. The SCA-35 is a masterpiece of compromise and optimization, that worked very well in thousands of builds and one can not make changes without a good understanding of how it works in synergy.
The RCA datasheet has a “characteristic class A working point” with a screen voltage of 50 V. I am sure David Hafler and the Dynaco carefully evaluated their operating point and chose one that worked best. Don’t forget that a change in screen voltage dramatically changes the pentode parameters, and with direct coupling between the pentode and triode sections it could upset the whole operation.
Unless you are true expert I suggest that you undo all the “unauthorised improvements” and rebuild your PC-10 exactly as designed with good parts. The SCA-35 is a masterpiece of compromise and optimization, that worked very well in thousands of builds and one can not make changes without a good understanding of how it works in synergy.
Last edited:
Right now, I am using 6BL8 with the mod mentioned in the post #16. Here are the voltages.
Pentode:
Plate: 83V
Grid No.2: 60V
Cathode: 1.2V
Triode:
Plate: 235V
Cathode: 94V
I didn't come up the mod from the thin air. I happen to have Dynaco ST-70, too. It doesn't suffer the issue that the SCA-35 has even with the same tubes. If you look at the schematic of ST-70 below, besides biasing through the cathode resistor, it also has 330K from the supply voltage to the cathode. It allows to use a lower value cathode resistor to keep voltage gain high. My idea is simple, migrate the bias scheme from ST-70 to SCA-35. Personally, I would avoid the Grid Leak bias with 4.7M. You don't really know how much leak current coming out of a tube. It could be anywhere. That's why I want to bypass C17 in SCA-35.
PS:
Also, ST-70 has the RC network in front of the phase spilter just as the Dave's design. I don't really agree with that the network is necessary for ST-70. However, I believe it is required for SCA-35 to get stable with wide range of tubes other than 7199, due to the existence of the possitive feedback.
Last edited:
