CF driver Va - Vk voltages in SSS consensus?

[ElectronAvalanche]

and the rest

[Richard1001]

Hi Rootz,

interesting analysis you did with the Fusion and Eagle technique regarding the similarities between some ODRs and SSS.
Looking at the power supply of SSS002 (also briefly touched upon by Richard re the Standby switch) I can not comprehend why there is a first filter (2x47uF in series) followed by the Standby switch then by two 100uF in series. Usually one would expect the larger filter node to come first in line. So in essence: 2x100uF in series, then Standby followed tby the choke then into the subsequent filter nodes. The setup as in 002 does not make sense and there is quite the current inrush into the 2x100uF in series (50uF in sum) once the standby switch is switched. So in essence you say that either the 2x100uF or the 2x47 do not really fit into the 002 chassis (we also do not have a pic of the power supply of 002). So could this be a mistake in the hand-drawn schematic? Could the 2x47uF be part of a node for the CF driver somewhere on the side of the chassis?

Here is a pic of ODR 03x (in the low thirties) I took myself. Two 100uF (do not know if in series, I see no balance resistors) followed by three 22uF. The other pic shows a three-section can cap. Unfortunately I did not take a general view pic, or at least I can not find it. AFAIR the preamp tubes were all in a row on the back, 2 power tubes (50W amp) with holes for two more output tubes. So essentially a chassis that could fit either a SSS or a ODR.

I will look for some more pics.

So again the question: is there also a mistake in the Japanese schematic regarding the power supply for the CF driver? Is there a dedicated power supply node for the CF? How could this look like? Would this node simoply have a way lower B+? Yet the bias supply seems to still have a high neg voltage derived by tapping the B+winding as in the Macintosh amp. So no real bipolar circuit in 002. So many questions. Maybe John Mayer would be so kind as to have his tech take a better pic of 002 next time they service the amp.

I need to upload the whole lot of pics I have someday. I had some pics posted here many years ago.

Electron

Smaller filtercaps where used quite often first in powersupply's that combine choke input filtering with capacitor input filtering in an attempt to get the best of both. Mostly seen in a tube rectifier powersupply

On the Winterland gutshot, the exact same thing can be seen. After the rectifier there are two smaller capacitors in series. These are mounted on a board to the far side panel of the amp chassis. After these caps there is an input choke. The output of the choke goes to the powersupply board as we know it, with the two big and three small caps.

So the choke could be drawn wrong, but this is unlikely because the Japanese schematic states it is a Fender style choke. This choke could not handle the high current needed.

The second option for the two 47uf series caps is that the powersupply can not be turned off completely. I found that as the Japanese schematic is drawn, schutting off the high voltage completely results in the powersupply caps being loaded with a negative voltage from the negative bias supply trough the CF tube.

By putting a resistor over the standby switch, the high voltage lowers, but does not shut off. The power caps can't get loaded with a negative voltage anymore. As a result the anode and screen voltage will go down and the voltage on the grids will get more negative. Shutting off the power tubes. I have not tried this in my amp, but putting a resistor over the standby switch has been done in other amps.

[martin manning]

That rectifier -> cap -> Standby Sw -> Cap is odd, but I think whomever traced it out would be unlikely to get that wrong; it's not complicated.

[Richard1001]

No it is not complicated and i don't think it's wrong, just a resistor that was missed. The more i think about it, the more i believe the high voltage is not supposed to be shut of completely for the reasons i mentioned earlier. Other simple things also got drawn wrong. It happens when tracing a circuit. The strangest mistake to me was the 1n bright cap over the MV wired to the mid switch. This should have been a 1uF cap wired to the phase inverter.

[rootz]

Wouldn't a reverse biased diode over the screens supply cap do the same? I mean, how much current would flow? 4mA?

Based on a couple of gut shots I made a schematic of the PSU of the 150W SSS005 clones I know. I think it looks like this. The choke would limit the inrush current a bit. Please do note: the resistor over the standby switch and the diode are NOT in the gut shots I have. That is me tinkering with suggestions to mitigate the negative voltage problem.

[erwin_ve]

Wouldn't a reverse biased diode over the screens supply cap do the same? I mean, how much current would flow? 4mA?

When building the Sss002 in 2020 Richard came up with that idea too. It is implemented in my Sss002 filter board.

[rccolgan]

Fascinating stuff to read here, guys. I'm most excited that there is continued evolution and discussion around this circuit

If it helps the discussion, the SSS 002 PCB prototype I built with Erwin's boards (Dumble CF) is still running strong today on the original TAD 7025 high grade tube I built it with. 30+ gigs and countless of practice hours. I built it for my friend that's local because I knew he'd beat the sh*t out of it with his SRV stuff and I can report here any findings in component selection, circuit design. Etc. I have not seen the amp in 3+ years for any service *knock on wood*

That said, I'm open to better approaches to the design and I'm intrigued by the bipolar supply idea. Very cool stuff! That CF tube does get hot in the original!

[erwin_ve]

Fascinating stuff to read here, guys. I'm most excited that there is continued evolution and discussion around this circuit

If it helps the discussion, the SSS 002 PCB prototype I built with Erwin's boards (Dumble CF) is still running strong today on the original TAD 7025 high grade tube I built it with. 30+ gigs and countless of practice hours. I built it for my friend that's local because I knew he'd beat the sh*t out of it with his SRV stuff and I can report here any findings in component selection, circuit design. Etc. I have not seen the amp in 3+ years for any service *knock on wood*

That said, I'm open to better approaches to the design and I'm intrigued by the bipolar supply idea. Very cool stuff! That CF tube does get hot in the original!

I also still have the "hot"CF tube in, remembering going through the spec sheets of different 12ax7/7025 tubes and the TAD had the best specs for that spot. That was 2020, today's TAD are different for sure.
As soon as the CF gives up I will put the bipolar in position.

[martin manning]

That CF tube does get hot in the original!

Plate dissipation in the original CF driver circuit is 700 mW at idle, where the bipolar is 175. 12AX7 max dissipation is 1W, so 700m is still within that, but most preamp tubes run something like 1 mA and 200V, or 200 mW.

PS The bipolar supply running from the 60V bias tap concept is now 7 years old: https://ampgarage.com/forum/viewtopic.php?f=4&t=29882

[JD0x0]

for any output tube cathode follower directly coupled, +150 -100V will be enough blind guess, also mosfet can easily work there

But with the MOSFET, you won't get the unique DC coupled CF 'distortion' achievable with the tube. In some cases that may be desirable, but it's worth noting.

A note about CF tube failures. This happens with some Marshall circuits. Apparently some of the Russian made tubes have more fragile cathodes and can't handle full 180VDC heater-cathode that 12AX7's should be rated for. I've been using NOS 12AZ7's. Basically a 12AT7 with a higher heater-cathode rating. They give a lower output impedance than 12AX7's too. Apparently, the Chinese 12AX7's hold up well, too.

[martin manning]

A note about CF tube failures. This happens with some Marshall circuits. Apparently some of the Russian made tubes have more fragile cathodes and can't handle full 180VDC heater-cathode that 12AX7's should be rated for.

I agree. This is probably the main issue with the SSS CF failures. The original design [edit] starts up at -225V Vk.

[bepone]

But with the MOSFET, you won't get the unique DC coupled CF 'distortion' achievable with the tube. In some cases that may be desirable, but it's worth noting.

there is no CF coupled distortion? tube working like follower and is just passing the signal with almost unity gain and 100% feedback.
what kind of distortion did you measure in this application?

[rootz]

A note about CF tube failures. This happens with some Marshall circuits. Apparently some of the Russian made tubes have more fragile cathodes and can't handle full 180VDC heater-cathode that 12AX7's should be rated for.

I agree. This is probably the main issue with the SSS CF failures. The original design idles at over 220V Vk.

I never quite understood why the voltages were so low in the hand drawn schematic Martin. Those cathode voltages are a bit safer, though there still is no protection for the CF's.

But, would it be possible that 002 was built with SF Twin reverb transformers and B+ was actually more around 405VDC?

Then again, it is super easy to add protection diodes over the preamp CF's. I'd just do that anyway. Adding an elevated heater supply is about as easy. Simple improvements for a more reliable amp. I probably couldn't explain a customer with NOS Telefunken EEC83's that they should hold up and get away with it when they do blow up.

[martin manning]

In either case Vg-k doesn't get out of hand, so I don't think there is much danger of arcing there. Low voltages on the Japanese schematic are possibly just due to Japan's 100 VAC line voltage. The 378V at the plate node would become 454V with 120V input.

[WhopperPlate]

But with the MOSFET, you won't get the unique DC coupled CF 'distortion' achievable with the tube. In some cases that may be desirable, but it's worth noting.

there is no CF coupled distortion? tube working like follower and is just passing the signal with almost unity gain and 100% feedback.
what kind of distortion did you measure in this application?

DCCF like in Marshall isn’t what we have in SSS2 phase inverter driver stage … but they are there before that…