Discussion in 'Tube Audio' started by GordonW, Jan 16, 2009.
Photos would be helpful here, Gordon. :yes:
I've taken a few photos of the bottom and top of the unit I'm working on. I'll take more, if needed, once it's done.
Progress on this thing is just really lurching along in fits and spurts. FINALLY got the line stage working correctly. Then, I discovered a number of out-of-spec resistors on the power amp board- which was part of the early-clipping problem. That, and a pair of 12AT7s subbed for the 12AX7s in the line-stage section on the preamp board, and I now have CLEAN AUDIO from a line source. Sounds GREAT.
However- to counter that, now the phono stage is only outputting on one channel. I know the problem is before the selector switch, in the phono stage itself. It actually was playing on both channels- until I swapped out the 12AX7s. At that point, something in the board must have cracked at a connection. Time to start continuity-checking and voltage-checking everything around the EF86 on that channel, and go from there. Two steps forward, one step back, once again...
Word to the wise- DO NOT disassemble one of these boards from the chassis, unless you NEVER want to use it again. They do NOT like being moved around/flexed. Any replacement of parts should be done with the board FIRMLY supported in the chassis (bolted in on all sides)... this will avoid cracking traces. Also, be SURE to NOT overheat the traces- part of the problem I had here, was ham-handed soldering by the original kit-assembler- many of the traces had started lifting. I've bridged many of them (by soldering along the trace itself, and/or bending over the component lead, to go along the trace further than the solder pad itself), and glued down several of the larger caps, to make sure things will stay secure...
Gordon, sorry this has been such a mess for you .Looking forward to getting it back. JOHN
AA-100 Restoration With Pictures
OK. Here's John's AA-100, after restoration:
Chassis bottom shot:
From this, you can see the basic layout... amp board at the bottom left, power supply bottom right, and preamp board at the top. You can see the new bias pots (blue), that were added.
Here's another picture of the same, with some of the more major work annotated:
John's unit was supplied with a 5R4 rectifier (more voltage drop then 5AR4/GZ34), so the big 80 ohm 30 watt resistor wound up being not being permanently installed- it was not necessary. If you're wanting to use a 5AR4 (necessary if you want to use the unit with the lid on it- the 5R4 will NOT fit under the factory cover!), then some sort of dropping resistor is needed to get the B+ down to a reasonable level...
Here's a close-up of the amp board- I circled the bias-test points (cathode connections of the output tubes) with a black Sharpie, to make them more apparent:
You can see the tiny 10 ohm, 1/4 watt resistors (BTW, one of them is brown, the other three blue- that was how the batch of resistors I got, came- same brand, same model, just one of a different color- go figure...), next to the black-circled terminals. On this board, the traces were cut to allow the output cathode connections to have individual grounds (a couple of jumper wires were needed, to run those new grounds to the cathodes), and the traces were also cut to allow separate access to each grid-bias resistor on the four output tubes- individual wires were run from each, to the bias supply pots.
One other problem that has been apparent on both AA-100s I've rebuilt, is that the screen supply resistors for the 7199s (820K ohm) tend to get "cooked" by the heat from the 22K phase-splitter resistors near them, when the coupling caps to the output tubes short. When this happens, these resistors go way out-of-spec... too high of a resistance value (typically 2M ohms or more, instead of 820K!) When this happens, the amp output becomes very distorted and weak. These resistors, as well as the 22K phase-splitter resistors, were all replaced. All the coupling caps were also replaced!
Here's a close-up of the power supply:
You can see the four new bias pots (labeled to correspond to the tube numbering on the output circuit board), mounted on a new terminal strip, and said terminal strip mounted to a small bracket which is then mounted to one of the transformer mount-down screws. This bracket moved the mounting point of the terminal strip rearward, so that the strip no longer ran into the wiring for the heater hum pots.
Each pot is labeled with its corresponding tube number, as silk-screened on the top circuit board. I also etched in the corresponding tube number on the bottom of the board, by the 10 ohm bias resistor for each output tube, so it would be easy to determine which test point corresponded to each bias pot.
Also, note the extra decoupling caps (the small black caps) near the bias pots. This was to isolate the tops and bottoms of the pots from each other, so that there was no "crosstalk" between them. Without those, the 8.2K resistor between the pots and the bias voltage source, and the 3.3K resistor between the pots and ground, would allow AC voltage drop across them... which would allow signal from one grid to "leak" to another, reducing gain and channel separation! The caps "shorted" such "AC signal leaks" to ground... no more crosstalk.
In addition, note that in this shot, the 80 ohm, 30 watt resistor is still in circuit. It's connected between the rectifier and the first power supply cap (60uf stock, 70uf total in this amp now). If using a 5AR4 rectifier, this cap is needed since B+ voltage, once the bias current through the outputs is reduced to a sensible (less than 35ma per output tube) level, and with today's high wall line voltage (commonly 120V or more, as opposed to 115-117V in the past), will commonly be close to 500V! This resistor drops between 15-20v... a value between 80 and 120 ohms is usually the right ballpark for this. But, since John had a 5R4 rectifier tube, this resistor was not needed... and it was removed, and the rectifier returned to being wired directly to the first power supply cap.
Lastly, here's a picture of the preamp circuit board:
In the earlier picture above, I noted that there were broken connections beneath the preamp board. These included several breaks in where the "auxiliary circuit cards" containing the selector switches (the little boards sticking up off the main board, with the shafts connecting them) connected to the main board! I had to go in, and solder in little jumpers, to re-make these connections, before I could get any useful output from the unit at all, on either channel!
Also, the tube positions on the left of the picture (V5 and V6 by the factory labeling) are the most important places where replacing the 12AX7 factory tubes, with 12AT7 or 6201 tubes, makes the most improvement. There can be a small improvement in using them in the phono stage 12AX7 positions as well, but the line stage is the biggest thing that makes a difference, when swapping tube types here.
One problem I did run into on this unit, that I did not on my earlier subject, was that the 2.2M ohm screen-supply resistors for the EF86 tubes, had both gone effectively OPEN CIRCUIT. Strangely enough, the unit still passed sound- just with much lower gain than normal, and a "muffled" character to the sound (like there was a high-filter blocking everything above about 5KHz or so). However, once these resistors were replaced, it's now dynamic and tonally correct- plenty of highs and good gain.
Frankly, this particular AA-100 had more problems than any of these I had encountered before. Between an original kit-assembler who wasn't careful about how they soldered (too much heat and not enough solder!), and the fact that this unit has been dis-assembled and re-assembled a number of times... the boards had really taken a beating. I could understand someone wanting to throw out those boards, if they went through what I went through on this one! However, it does sound FANTASTIC now, once it's properly repaired and restored... and I'd expect that MOST of these units won't have NEARLY this many problem areas...
Thanks for sharing! I have one of these waiting on the shelf - maybe next year I'll have time to start it.
Really very nice, Gordon. Gives one some hope that these things may actually have more utility value than merely as iron mines to be pillaged and tossed.
Thanks for all the extra work you put forth taking photos, writing details, etc. I decided to mount the pots on a perf board so mine will end up looking a little different (but electrically the same), will post some photos when I complete it. What value did you use for those bias resistor decoupling caps? Thanks.
Now I am really looking forward to getting it back.
ok let me edit this
i found a place for cheaper caps YAY and i should order them in 2 months YAY !!!
You should. :thmbsp:
BTW: John, this applies to you, as well as anyone else with one of these- if you're willing to run the amp with the cover removed, you might want to find a 5T4 (metal-can, military type) rectifier tube. The 5R4 will work OK in these, but it drops B+ to below stock value (about 405v, in this case). I'd guess that the 5T4, being about half-way between the voltage drop of the 5R4 and the 5AR4 (original- which gives about 465v, on today's line voltages), would get you right in the ballpark (425-435v at the plates with the 5T4).
John, if you got a 5T4, you could probably just drop it in your amp, as-is. You'd want to check the test-points on the bottom to make sure you didn't go over 32ma plate current (.32v on a meter, between ground and the test points). 30ma (.30v on the meter) would make me feel better- but, it's easily adjusted by the pots. 30ma at about 430v is about 13w dissipation... about 68% of max for the tube. That's a good place.
AA-100, done for now
I am done (for now). I mounted the pots on a small Radio Shack perf board and attached it to a custom L-bracket which I fashioned from some scrap metal. I wasn't comfortable with the thin beat-up stock power cord, so I went all out and installed an IEC connector for a 3-wire grounded power cord. I replaced the line bypass capacitor with a new X1/Y2 safety capacitor, and grounded the chassis via the IEC connector.
I installed the 10 ohm cathode resistors directly on the PC board, and installed 220K grid resistors and replaced the output coupling caps, as well as most other coupling caps. I installed a pair of 120 ohm 5 watt resistors (paralleled for 60 ohms) to reduce the B+ voltage down to 455vdc. I have a 50 ohm chassis mount 25 watt on order, and will install it in place of the 2 120 ohm resistors. I am impressed with the sound of this amp, now it sounds really sweet and pristene, like a high-end tube amp. I have attached pictures.
Yes, thanks to GordonW for those mods. I really like the sound of mine too...
Thank You so much Gordon, I've had two of these amps sitting in storage for decades awaiting their wake-up call, (I bought one with the matching AM/FM tuner, found the other one cheap later and thought having extra parts wasn't a bad idea).
They're currently getting rebuilt and I sent your most generous information to him, and he came back with happy news, he can fix them both! So, now I have one AA100 to spare!
(now the eternal dilemma, sell it or find a second tuner...LOL)
Thanks again for your generosity.
I am new to tubes and specifically bought this amp because of this thread. I thought what better way to learn then by doing the work. I spent some time researching tube theory then I did Gordon's mods and the amp sounds fantastic...with an MP3 player as a source. I now have a CD player on the way!! I am hooked and am looking forward to refurbishing an SA-2 next. My dad had one when I was growing up that he had built as a high school student. I hope to have it working by his birthday so I can give it to him as a gift. Thanks Gordon!!
Hey Jared, Read this thread. JOHN
Thanks John for bumping the thread for me. Hey Gordon, sent ya a PM.
I recently got an AA100, all original tubes - 2 of the 7591A redplated after about 10 minutes. After reading your excellent post, recapped 6 of the 0.1uF 400V coupling capacitors in the Power Amplifier PCB with Solen metallized polypropylene capacitors.
After recapping 7591A tubes did not redplate. Grid Bias now reads -18.2 on all 4 tubes.
The left channel now does not seem to have same power and bass response as the right channel (left sounds weaker). Before recapping they were OK - there was no perceivable difference untill the tube redplated (10-20 minutes). Should I replace these Solen coupling capacitors with a different brand like AuriCap?. Has anyone tried Claritycap? They seem to have tighter tolerance rating +/-3%
I am planning to modify the tube bias according to your diagram probably tonight.
Good caps are good caps even prosaic Panasonic or Sprague metallized polyprops, those shouldn't make that difference. More likely tube condition issues if all resistors have been vetted during refurbishment, red-plated tubes are overstressed tubes.
I got a couple messages about these amps recently, so I thought I'd reiterate a couple things that keep coming up.
First off- again, be CAREFUL with these circuit boards. They can be somewhat fragile. If you even THINK a connection could be weak, it's a good idea to carefully (without burning up things) flow some extra solder back from the connection, a good way back onto the solder trace, to strengthen it. Also, using a spot of glue to hold down heavier components (like caps) once they are soldered in properly (not a lot- just a spot, to keep them from vibrating around) is a good idea. You don't want to prevent someone in the future from being able to change out a cap should it go bad, but you do want it to not move around, and stress the solder joints.
Secondly- out-of-tolerance resistors on the power amp board seem to be RAMPANT on these. Especially the 820K screen resistor on the pentode side of the 7199, and the "top" 22K phase-inverter B+ resistor on the triode side of the 7199. In fact, if your AA100 has 1/2 watt 22K resistors there in the phase inverter (some do from the factory), it's a good idea to replace both of them (two per channel) with 22K 1 watt carbon-film or carbon-composition resistors, for additional power handling.
I concur, I had to add some wires to connect several points on the boards (mine were toasted, particularly around the power tubes).
Separate names with a comma.