SS Rectified Get*Set*Go

[lth1]

I have been listening to the GSG for the last couple of days and was not at ease with the sound that I was hearing, something to me was just not right. Remembered a comment by Ed Brown on using UF4007's in an octal socket for SS rectification. Only had one UF4007, but had alot of UF5408's, so I built Ed's socket.

The Sound to me is alot better from top to bottom, a simple cheap experiment.

Have also been listening to the 6B46's and 6A3's and their isn't enought difference between the two brands that I have to recommend one over the other.

Lee Hankins

[Shannon Parks]

The 6AX5GT has a large voltage drop with these currents - maybe 45V - but I kind of wanted to run the NOS tubes conservatively, and this keeps the max plate voltages in spec. Certainly, using some SS diodes will dramatically raise the B+ and hot rod the amp some. It'll help squeeze every last drop out of the DHTs, but it might not be recommended for some NOS tubes.

[EWBrown]

If you run with the resulting higher B+ (around 360VDC) I would suggest changing the 6B4G cathode resistors back to the originally specified 1100 ohms, just to ensure longer tube life. Or, perhaps, try both values , then stick with whatever RK value sounds better.

On my first build, I started out with three 3,300 ohm 3W WW rsistors in parallel, for 1,100 ohms, then when Shannon suggested lowering the cathodes' resistors to 900 ohms, I connected a 4,700 ohm 3W WW resistor in parallel, and that got it pretty close to 900 ohms, without performing "major surgery".

I later got some nice 1% 3W 2.74K resistors from Hand Made Electronics and replaced with three in parallel for each channel.

/ed B in NH

[lth1]

Please do not try this without aditional information, I blew several caps (C5 & C6) and possibly the power xfmr. This project is on hold for awhile.

Lee Hankins

[EWBrown]

If the "cathode" bypass caps C5 and C6 blew out, that would indicate that the voltage across them was excessive (over 100V). That would be caused by excessive plate current, or perhaps the cathode resistors' overheating and failing out of spec, with increased resistance, hence increased voltage drop.

The 6B4Gs' plate-to-filament voltage shouldn't exceed 300VDC by too much margin (the original RCA 2A3, 6A3 and 6B4G specs were 250VDC@ 60 mA), I've seen 320VDC P-K in some designs
but the current was kept lower by using 1000 to 1200 ohms "cathode" resistance, or by using approximately -60V grid biasing, in this case by using the aforementioned "split rail" power supply technique .

I checked the Paramount II pdf Manual, it indicated that the "split rail" PSUivers +320VDC to the B+, and -60 VDC to the nagative rail, and ctual measurements were fairly close to those. The optupt stage design is "parallel fed" in that it uses a separate plate loading choke, which is capacitor coupled to the OPT (a 4K primary, 4 and 8 ohm secondary line transformer). The plate chole in this case is a Triad C-7X, 10H, 90 mA and DCR of 270 ohms. In this circuit, the B+ is SS rectified.

My own measurements indicated about 54 mA plate current (59.5V across the 1100 ohms, 10W WW "negative rail" resistor) so that the measured 325VDC B+ drops about 15VDC through the choke, so the plate voltage is around 310VDC, with the filament connected to the ground / zero volts bus, via the usual 50 ohm "hum" pot wiper to ground.

The Sovtek / Elektro-Harmonix 2A3s, 6A3s and 6B4Gs are basically identical except for filament voltatges and basing, the plate and grid structures are the same. They can safely take a lot more "abuse" than the vintage 1940s-1950s "gassware" could handle.

/ed B in NH

[Shannon Parks]

Sorry to hear this, Lee. Don't worry about another GSG board, Lee, as I'll ship one out gratis today and refund your Paypal.

C5 & C6 are the cathode caps and are rated at 100V, so the power tubes must have arced over and consequently put ~300V+ on them, blowing them. The power tubes are probably damaged but I would guess the power tranny is OK. I would replace C5 & C6, and test the tubes in a tube tester. Go ahead and revert to the 6AX5GT and test out the power supply sans tubes, to see if the power transformer is OK. Of course you can always test the AC voltages unloaded, too. Did a fuse blow perchance?

FWIW, the modern 2A3-variant DHT tubes can handle a much higher B+ than their ancestors. Earliest datasheets were at 250V max plate and later at 275V. The Sovteks can handle 450V and can be run like a raped ape. But I think the original design and now age of the old DHTs requires some gentle circuits. I've had one 'sparkle arc' in my Hickok, even!

[lth1]

Thank you Ed and Shannon.

Shannon the error was due to my ignorance and I do want to build another GSG just do not want to mess with the existing board at this time; so please do not refund the cost of the new board. I have also purchased a new power xfmr and James output xfmr's, will be interesting to compare the sound between the MagneQuest and James output xfmr's.

Have reconnected the system with my Bottlehead paramours, the "HOT" SS rectified GSG with 6A3's sounded much better than the Paramours, but not with tube rectification, there seemed to be an upper frequency distortion to the sound with the 6AX5gt, this distortion was absent with SS rectification.

After replacing the cathode caps and installing the 6AX5's, the unit would play for several minutes then the rectifier would arc and blow a fuse, this happened with both tubes that I have. There was also a lot of popping noise which seemed to be coming from the power xfmr. Just checked the power xfmr with no load and it seems to be OK, but found another cathode caps rather bloated, did not go off as violently as the first ones did.

[EWBrown]

It looks like the (6A3) tubes are internally arcing over, or at least drawing excessive current, and this would raise the voltage (and current) through the cathode resistors and the C5 and C6 caps, which would "bloat" them. Also, just for a sanity check, make sure that C5 and C6 are oriented correctly (I've installed caps backwards more times than I care to admit) Yellow_Light_Colorz_PDT_02 .

Check the total cathode resistance for each side, with the power off, all PSU caps discharged, and tubes out of the sockets, and verify that it is still the correct value.

While all the tubes ar pulled out, just power it up and let it idle, to verify that the power trannie is OK, with no load on it. The measured AC voltages with no load will be around 5 to 10% higher than under normal operating conditions, this is normal. The transformer may warm up a little bit, but should not become hot to the touch.

HTH

/ed B in NH

[lth1]

Going to try to post for the third time today. My computer is having a melt down just like the GSG.

Ed, I did as you recommended and everything seems to be OK, except the two 6AX5's which are dead. I need to clarify some previous information, C6 is the only cap that has gone bad, it exploded the first time and did the "bloat" thing the second.

The cathode resistors are again 3 X 3.3K for 1100 ohms. I had removed one 3.3K and replaced it with a 2k 5W for 890 ohms. Replaced C5 & C6 with a JJ 100/100uf 500V cap, it doesn't begin to get warm with SS rectification.

I listened to the GSG last night and this mourning (checking the JJ caps temp every 5 minutes), and much prefer SS rectification. Ed how do I go about doing this.

Thank you ahead of time and for all previous help and advise.

Lee Hankins

[Shannon Parks]

Hi Lee,

I've hacked up my GSG to a hybrid biased 300B, so I can't easily recreate your setup. Could you take the voltages at the plate and cathode of the 6A3s? I'll try and determine if they are biased within the limits. Earlier I was thinking the max plate voltage was what caused the tubes to arc over, but it was probably just good 'ol thermal runaway with an over biased tube when the 900 ohm network was in there with the increased B+.

Good luck with the computer probs, too!

[lth1]

Hi Shannon,

The plate voltage is 387v and cathode is 60.4v and 66.7v. I tried your 300B conversion today but only had a 2 ohm 50w resistor for R19 & R20, would this cause a problem.

Ed in a previous statement that SS rectification would increase the voltage by 45v, in my SS GSG it increased by 75V's +, just an observation. Made two different SS rectifiers one with UF4007 the other with UF5408 all measurements were identical, didn't expect a difference but just had to try.

The first cap that violently let go was with the 1100 ohm network, with the 900 ohm network the cap shorted and "bloated". The 6AX5's both shorted and blew the fuse with an 890 ohm network.

Lee

[SDS-PAGE]

Since we are talking about SS retification, has anyone tried using one of Weber's copper cap retifiers? The WAX5 is a direct drop-in replacement for 6AX5.
http://www.webervst.com/ccap.html

[Shannon Parks]

Hi Shannon,

The plate voltage is 387v and cathode is 60.4v and 66.7v. I tried your 300B conversion today but only had a 2 ohm 50w resistor for R19 & R20, would this cause a problem.

Ed in a previous statement that SS rectification would increase the voltage by 45v, in my SS GSG it increased by 75V's +, just an observation. Made two different SS rectifiers one with UF4007 the other with UF5408 all measurements were identical, didn't expect a difference but just had to try.

The first cap that violently let go was with the 1100 ohm network, with the 900 ohm network the cap shorted and "bloated". The 6AX5's both shorted and blew the fuse with an 890 ohm network.

Lee

Ok so one tube is biased at 55mA and the other at 61mA, for 17.9W and 19.5W dissipated at idle, respectively. Still too hot, and I definitely see the thermal runaway at 900 ohms. Let's try and get this under 15W. But the voltage drop across the tube isn't bad - I see that a 6A3 is rated at 325, like later 2A3/6B4s. So we should be cool with the SS B+ after all. You can take one of your old 2.7k, 3W resistors and put it in series with R15A and another at R16A. Should increase the newtork to 1294 ohms. See if that gets us under 15W.

[lth1]

Hello Shannon,

I ordered a Weber copper cap retifiers, will change everything back to stock (1100 ohms network) and see what happens. If it does exactly replace the 6AX5 then everything should be OK, hope that the B+ delay will eliminate the present pop on startup (SS rectification only).

I know that I did not give you the correct cathode voltage, please tell me how to measure this correctly. Secondly, what are the test point for, their was no mention or instructions in the print out. I am sorry for my complete lack of knowledge in this area.

I did listen to the GSG with 300B last night (with R19/20 at 2 ohms instead of the recommended 1.5 ohms) and I do not believe that it sounds as good as the 6A3 tube. But of course everything isn't exactly running at proper specs and not enought time has been given for a proper comparision, not an unusual conclusion for one who prefers the 2A3 over the 300B.

[EWBrown]

I found that the two Sovtek 6A3s I got from TUbe Depot were not very closely matched, one is a bit "greedier" than the other, and on a quick emissions test, one scored "80" and the other "68". BTW the Sovtek 6B4Gs were almost as as disparate. Both of these are the "mono-plate" version of the tube,

I just recieved the NOS Russian 6C4Cs from "sovcom". These are the old style "twin plate" version, with the black finish on the plates. Of the four, three were a near-perfect match, and the fourth one was just a little bit off. Not bad for "randomly" picked tubes - the seller never mentions testing or matching hs tubes.

Back to B+ voltage reduction techniques:

First way is to connect a wirewound power resistor in series with each SS rectifier between the power tranny secondary "hot" leads and the rectifier anodes. These will get hot, nd should be located where they can get some air space around them. If you used the terminal barrier strip, just place them between the tranny "red" leads and the appripriate connections on the barrier strip. THe resistors should have a rating of at least 3 times the actual dissipated power, so 10 watts is a bit too low, 15 to 20 watts (or higher) would be good. The previously mentioned "copper cap" rectifier already these resistors built-in.

Another way to reduce the B+ voltage with SS rectification, is to replace the electrolytic cap betweem the rectifiers and inductor with a lower value MKP (Solen Fast Cap or similar) start around 5 to 6 uF, 630VDC and work up from there. This lets the choke absorb some of the excess 100 / 120 Hz "ripple" and thusly generate lower DC voltage down stream.

The 1.5 H choke will have an AC reactance of around 1130 ohms at 120Hz and 940 ohms at 100 Hz.

The potential downside of using this approach is that the choke will have more of a ripple component across it, so it will generate more of an external magnetic field, and may generate a machanican vibration.

HTH

/ed B in NH