Discussion in 'Tube Audio' started by GordonW, Jan 16, 2009.
Gents - All this talk about using ten ohm resistors as fuses leads me to a better idea. How about putting REAL fuses in there! I always fuse the B+in my tube amps. Once the amp is running properly, turn it off, discharge it, pull one B+ lead and insert a DC current meter in series. Then run the amp to near clipping and read the current. Say it's 250 ma. I double it and use a 500ma, Fast-Blo fuse. Don't depend on a resistor to burn out, it's too slow a process. And don't depend on the mains fuse to protect anything either. It's there simply to prevent a fire from starting. The mains fuse has to be chosen to pass the high voltage currents and the heater currents as well. So it may not blow just because one tube red-plates. Repeat the process for the other amp channel if it's stereo.
A fuse doesn't allow you to sample the cathode current. You'd need a fuse AND a resistor, for EACH output tube.
Given that, a 1/4 watt 10 ohm resistor works better as a fuse, than nothing. It's saved transformers for me, in several cases. It will usually blow faster than even a global B+ fuse- because it's small enough to only handle the ONE tube it's connected to- in contrast to the B+ fuse, which has to handle ALL the output tubes. U
There's at least one Pioneer solid state receiver that uses a 1/8 watt resistor for the same purpose. The component number is FRn instead of just Rn for "fusible resistor".
Learning stuff here on AK is just pure fun.....
Any comments on this circuit design? It is from a gentleman that is selling replacement circuit boards for the AA-100. I have purchased and installed the board which is very nicely designed and documented. But in my case I cannot get the bias voltage to come down on 3 of my tubes. They are reading around 48mv with the 10k pot at minimum. He suggested getting the tubes tested which I will do shortly. But I am curious if anyone has an opinion on another possible solution. My tubes are original Daystrom tubes and I would love to keep them. Some notes: the 220k resistors R91,92,85,86 were supplied as 200k. The voltage going into P40 is -17 volts. My power supply is putting out 488vdc.
Weird. I can't download that PDF for some reason.
I've got a basket case aa-100 that I'd like to revive and I'd love to hear the answers to your question....
If someone is going to make new boards for the output amp- then I would suggest making them work with the 6GH8 for the driver tube, instead of the 7199. This has been done with many amps that used the 7199, with good success. The 6GH8 is pennies on the dollar compared to the 7199, and seems to work just as well.
Is this the board you've got? If so, have you done the bias fix? https://aa100restoration.wordpress.com/
Yes I did the bias fix. He included an instruction sheet for that process. I can adjust the bias upward but cannot get it down to the 33ma recommended.
-17v at P40 isn't enough. That should be at least -20v.
Is the bias rectifier a silicon diode (NOT the selenium rectifier anymore)? If it has not been changed to a silicon diode, then it NEEDS to be.
If the diode has already been replaced, and the 3.3K resistor (R106) is already gone (which seems to be the recommendation as part of the bias mod)- then replace the 8.2K series resistor (R103) in the bias supply, with something like a 6.8K. That should get you up to near -20v at P40. Then, you should be able to adjust the bias to a sensible value (less than 35ma per tube).
I have removed the R106 3.3k already. I will try adjusting the value of the R103 next time I have the soldering pencil heated up. I did test my tubes today. They are right on the line between good and bad. But I really liked the idea of having the original Daystrom tubes. Oh well, new circuit board, new tubes. I hope the 7199's are good. Those are getting very rare. I ordered a whole new matched quad of 7519a's just so this project will have a hope of sounding good in the end. I will put the others on the mantel as a trophy. But I don't believe that marginal tubes would affect the ability to adjust the bias. Right?
Marginal tubes should, usually, have LOWER current at the same bias negative voltage...you seem to have the opposite problem (too MUCH current).
What did you test the tubes on? Many of the "drug store testers" (the big stand-up ones that were in stores) are designed to "sand-bag" tubes. They were designed to sell tubes, so many of them would read low, so that a marginal tube would test already bad... so a tube that tests marginal on such a tester, may still be fine.
I went to the Pavek Broadcast Museum here in Minneapolis. They have a tube tester that you can use for the price of admission. Along with expert advice and great conversation. A whole room full of tube experts is pretty much priceless for a man in my situation. The tester reminded me of an Enigma machine. He showed me how to use it and I tested my whole box of tubes. I'm working on a W5M also. But the AA-100 has to sing first. Some day I will get back there and check out the whole collection. They sell tubes there also but did not have the ones I needed. So I went to AES.
Once you get the bias supply sorted out you should be able to bias your tubes okay. I also have this board and I can bias my mixed bag of 7591s correctly (after the bias fix).
This is correct. The supply needs to be at -20 volts to be effective. Have you replaced the resistors in the bias supply circuit? It's also possible that the selenium rectifier is getting weak so replacement may buy you some more volts.
That's a good idea Gordon. I originally made these boards because mine was thrashed and I needed a new one. I already had a good set of 7199s at the time so it wasn't an issue for me.
Since the topic came up, is there sufficient interest in a newer, better PCB design that has:
- 6GH8 capability
- Integrated bias supply
- or ??
What about a retrofit pre-amp board?
What do you guys need?
The first run of PCBs is nearly gone (thank you!). I would be glad to make more AA-100 parts if there are enough people who want them.
I do still have my original selenium rectifier in the circuit also. Any one want to throw out a part number for a silicon diode replacement?
I prefer UF4007 high-speed diodes for that. Though, a (ubiquitous) 1N4007 will work fine. Even the lower-voltage diodes such as the UF4004 or 1N4004 are just fine.
Replacing that should add a couple of volts to the supply. May be all the problem.
Ultra fast soft recovery rocks!
This is just speculation, but I suspect the participants here would enjoy having access to some sort of universally applicable PCB for this style of power amplifier.
I finally got my diode in the mail [along with some tubes for a W5M] and put it in. That raised the voltage up to -22vdc. And I could do my bias adjustments just fine.
So then I moved on to replacing all the capacitors on the preamp section. "Orange drops" for everything except a couple yellow spragues and a couple polarized caps. Then I put all the tubes back in and tried it out. All the crackles and pops were gone. The amp sounded better than ever before. Then I went back and replaced all of the resistors on the preamp. And just because I was getting tired of testing the quality of the sound I reworked the power supply section before giving it a try again. New pop can capacitors and all the other resistors and bits in the power supply section were replaced. When I powered it up again the right side sounded just fine again but the left side was pretty much a hum fest with a tiny little audio signal buried in there somewhere. I figured I had done something terribly wrong in the power supply section. Eventually I decided that the power supply and amplifier section were not the problem. The main clue was that if I shorted the A signal to the B signal I got almost good audio out of both sides. So I messed something up when installing resistors. Or the power supply took out a preamp tube or something. The power supply was going up to 550vdc for the B+ without any tubes installed. That is getting close to the limits on some of those components. The B+ usually is at 475 when the tubes are installed.
So I am going to go through the preamp section with a signal generator and oscilloscope (avoiding the high voltage bits) to see if I can find a place where the signal drops out. A visual inspection and reflowing the solder connections did not yield any results. I will probably try swapping tubes between the left and right sides to see if that reveals anything.
It's kind of nice working on a project that does not have a deadline. . .or occupancy date for those in the fire alarm trade.
Separate names with a comma.