Yeah I wouldn't have done what I did if I hadn't read the report straight from someone who works at ASC that said that it would be OK. I don't remember the actual values because I've been working on these amps for a long time though.kevinkr said:
I think they are too and that it is time for a nice quad of KT77's.kevinkr said:
I'll take a look.kevinkr said:
Sounds cool! I'll try to get a couple UF4007's shipped from home to try this. Of course, I have to do whichever modifications to both ampskevinkr said:
But I ask you, how do YOU know what DC ratings ASC has ascribed to their capacitors?anatech said:
I'll try to check on as many of these as possible. I'm just wondering, though, how I could "overheat" the tube if I've never seen redness on the plates. The screen grid maybe?anatech said:
At the risk of sounding curt, here goes. A) I never "removed" the choke or proposed to do any such thing. B) Replacing or disconnecting the caps, however, I did suggest.anatech said:
The whole schematic, PSU included, is actually in my first post
Hi Sorenj07,
Looking at your schematic, I am drawn to your grid supply. Capacitors can burn foil off a bit at a time. The capacitor then loses capacitance and your bias voltage may then drop. I am not comfortable with capacitive dropping devices. Please tell me you at least used a foil cap and not metalized.
Also, it may take some time to allow the bias voltage to reach it's intended value. Too bad you didn't have a bias winding.
One last thing to check. You probably don't have a issue with this, but do make sure your bias controls are not intermittent and don't arc to ground (more properly, have leakage resistance to ground).
-Chris
Thanks. I thought I saw it, but I was in a rush at the time.
1N4007's will be fine, believe it or not.
I would highly recommend you consider 6L6EH. Excellent tubes. I have also had problems with Sovtek tubes at high B+. They are not reliable at 450 VDC. Been there, done that. The 6L6EH are very reliable and sound great. Much personal and professional experience here.
Well, I don't have to know this. I only need to know what their rating is for one (the one printed on the case) and the conversion from there is easy. Any other verbal or emailed information is highly suspect. It is the listed rating that counts. So, if your caps are rated at 330 VAC, I would expect a DC rating of 466 VDC (rounding down) as Kevin mentioned earlier. This may be a cap breaking down on peaks. The first one I imagine after your rectifier. You may want to consider installing a resistor in series to damp any possible ringing.
That is certainly possible for the screen grid in your output tubes. The control grid may also have gotten too hot or there may be flakes from the base sitting in the structure. One never knows on an old tube that has been transported a few times. The other thing to consider is that low level redness can still cause out gassing. The way to check for this is to examine the plates with the amp running in a darkened room. You will then be able to see the plate running beyond it's limits by just a bit.
Looking at your schematic, I am drawn to your grid supply. Capacitors can burn foil off a bit at a time. The capacitor then loses capacitance and your bias voltage may then drop. I am not comfortable with capacitive dropping devices. Please tell me you at least used a foil cap and not metalized.
Also, it may take some time to allow the bias voltage to reach it's intended value. Too bad you didn't have a bias winding.
One last thing to check. You probably don't have a issue with this, but do make sure your bias controls are not intermittent and don't arc to ground (more properly, have leakage resistance to ground).
-Chris
thanks for the continued support I'll consider the 6L6EH's, especially since it seems that 6.6K is a "better" load for the 6L6 family. As for the capacitative dropping resistor, it is a standard type dipped cap at 1KV by Cornell Dubilier. Should I substitute something a bit more robust? Maybe as an alternative I should just pick up a tiny little 50V bias transformer? how much current do 6L6's take to bias, assuming no grid current?
This thread is (I think) the one that led me to my thinking on DC ratings. If worst comes to worst I can cut the 12-gauge ground bus wire and sub in another 80uF 370VAC cap for the B+, it's the exact same size.
I haven't checked the bias adjustment but I do remember from when I was biasing my amps for the first time after a couple of days of running them, that the current was very hard to measure - it kept jumping around and whatnot. It took a while for it to settle down at whatever voltage reading across the 10 ohm cathode resistors... in this case, .55V or so.
What would you recommend as the first course of action before trying to run the amps again? Buy some 6L6EH's? Get a multimeter and check resistances? Get a soldering iron and reflow a bunch of joints?
I'll consider the 6L6EH's, especially since it seems that 6.6K is a "better" load for the 6L6 family. As for the capacitative dropping resistor, it is a standard type dipped cap at 1KV by Cornell Dubilier. Should I substitute something a bit more robust? Maybe as an alternative I should just pick up a tiny little 50V bias transformer? how much current do 6L6's take to bias, assuming no grid current? This thread is (I think) the one that led me to my thinking on DC ratings. If worst comes to worst I can cut the 12-gauge ground bus wire and sub in another 80uF 370VAC cap for the B+, it's the exact same size.
I haven't checked the bias adjustment but I do remember from when I was biasing my amps for the first time after a couple of days of running them, that the current was very hard to measure - it kept jumping around and whatnot. It took a while for it to settle down at whatever voltage reading across the 10 ohm cathode resistors... in this case, .55V or so.
What would you recommend as the first course of action before trying to run the amps again? Buy some 6L6EH's? Get a multimeter and check resistances? Get a soldering iron and reflow a bunch of joints?
Hi Sorenj07,
The most valuable test instrument you have are your eyes connected to your brain.
-Chris
The most valuable test instrument you have are your eyes connected to your brain.
Nothing compared to your bias circuit. A 50 mA transformer would be fine with small filter caps.
It may be a film type. You want foil. Look it up to see exactly what it is.
Well, you do need to do some troubleshooting before spending a dime. So .......
This is step one. You need to know where your problems are firstly.
Probably at some point in time - soon. It wouldn't surprise me to see the originals are misbehaving. Do the dark room test first. When you get your meter, make sure the bias current is equal between the two and that it stays equal. At the same time you can eliminate your rectifier problems. You can add the diodes last. First, find the fault instead of just covering them up. Look up the rectifier tube on the net and read up on it.
-Chris
Hi Soren, Few things...
I'm worried that you are measuring 500V on a choke input supply with a 1000VCT! I get about 450 with a 1150VCT transformer in my vintage Williamson amps when running rated input voltage. Is the input cap making trouble perhaps?
When you say "after letting it drop back down a little too quickly after looking underneath it", does that you mean you banged it around a bit while turned on? That could potentially cause an internal short in the tube, no big deal. Another possibility is -- and maybe I'm looking at your amp wrong, but isn't the lead of that .22uF cap quite close to the 100uF metal cap body? Does it have any insulation? Maybe it bounced and briefly shorted against the can (scary!), nuking the rectifier? Just speculating. Frankly, I'd consider dumping both the .22uF's anyways -- they make me nervous.
In any case, dying 5U4GB's tend to flake out in just the manner you describe. Many of them jingle in operation, which can be normal if not arcing. I've had a LOT of them that arc blue on turn-on (even with choke input), and those that fail arc white inside and continued to glow for a bit. They are "done" once that happens. It's possible you had an unlucky one.
Finally, another thing that can help your amp is the addition of a power resistor on the b+ supply to draw some current when the tubes aren't conducting. This keeps the choke input a bit happier by giving it an initial load -- the 5U4 warms up almost instantly, but the other tubes take some time, and this smooths that out slightly. A 47k 10W power-wasting resistor is a classical solution to this (c/o Craftsmen 500a). Sucks down an extra ~10mA.
g'luck!
I'm worried that you are measuring 500V on a choke input supply with a 1000VCT! I get about 450 with a 1150VCT transformer in my vintage Williamson amps when running rated input voltage. Is the input cap making trouble perhaps?
When you say "after letting it drop back down a little too quickly after looking underneath it", does that you mean you banged it around a bit while turned on? That could potentially cause an internal short in the tube, no big deal. Another possibility is -- and maybe I'm looking at your amp wrong, but isn't the lead of that .22uF cap quite close to the 100uF metal cap body? Does it have any insulation? Maybe it bounced and briefly shorted against the can (scary!), nuking the rectifier? Just speculating. Frankly, I'd consider dumping both the .22uF's anyways -- they make me nervous.
In any case, dying 5U4GB's tend to flake out in just the manner you describe. Many of them jingle in operation, which can be normal if not arcing. I've had a LOT of them that arc blue on turn-on (even with choke input), and those that fail arc white inside and continued to glow for a bit. They are "done" once that happens. It's possible you had an unlucky one.
Finally, another thing that can help your amp is the addition of a power resistor on the b+ supply to draw some current when the tubes aren't conducting. This keeps the choke input a bit happier by giving it an initial load -- the 5U4 warms up almost instantly, but the other tubes take some time, and this smooths that out slightly. A 47k 10W power-wasting resistor is a classical solution to this (c/o Craftsmen 500a). Sucks down an extra ~10mA.
g'luck!
Thanks! I'm considering dumping the .22uF caps but I'm a bit worried about the voltage spikes that the rectifiers (which will be 5R4GYA's or somesuch) will have to deal with. I guess I'll try an RC across the choke instead. However, the two caps are pretty firm in their position, even though they're flying supportless.
I'm still wondering how I could have 500V on the B+ as well. I expect in any case that the 5R4 will drop a fair bit more volts so that should help. After a lengthy discussion with Eric Hayes from DiyTube, I'm also probably going to change the layout of the amp while I've got it on the "bench" to something a bit more stable, a high-gain single triode feeding into an LTP with 6SN7's. I'm also considering making one of those long-tailed totem-pole pairs out of 6SN7's or perhaps 6SL7's for more gain..
In any case I have a fresh quad of JJ KT77's on the way to use when the amp is up and running again. The recommended/design center plate+screen dissipation is 28W, which is right where I had my 6L6's earlier, through no fault of my own (500VB+ was NOT expected
)
I'll do all the usual resistor checks etc. while running the amp with SS rectification hard-wired into the tube socket for now.
I'm still wondering how I could have 500V on the B+ as well. I expect in any case that the 5R4 will drop a fair bit more volts so that should help. After a lengthy discussion with Eric Hayes from DiyTube, I'm also probably going to change the layout of the amp while I've got it on the "bench" to something a bit more stable, a high-gain single triode feeding into an LTP with 6SN7's. I'm also considering making one of those long-tailed totem-pole pairs out of 6SN7's or perhaps 6SL7's for more gain..
In any case I have a fresh quad of JJ KT77's on the way to use when the amp is up and running again. The recommended/design center plate+screen dissipation is 28W, which is right where I had my 6L6's earlier, through no fault of my own (500VB+ was NOT expected
)I'll do all the usual resistor checks etc. while running the amp with SS rectification hard-wired into the tube socket for now.
Update - some good news, some weird. I put the other pair of 6L6WXT+'s into the right channel amp and they exhibited similar darker purplish glow as opposed to blue glow (this is on the envelope) and a bit of cloudy blue glow inside the plate structure.
Looking forward to experimenting with some '77s. 600V plate + grid max voltage is encouraging, and the extra filament current might help the heater voltage drop to a more reasonable level. While one of the amps is in the "sick bay" it's a good excuse to tweak the topology of the amp, in this case with a 6SF5 input tube and 6SN7 LTP.
Looking forward to experimenting with some '77s. 600V plate + grid max voltage is encouraging, and the extra filament current might help the heater voltage drop to a more reasonable level. While one of the amps is in the "sick bay" it's a good excuse to tweak the topology of the amp, in this case with a 6SF5 input tube and 6SN7 LTP.
Just found out the problem with my amp, now that I have a DMM. My negative bias supply is shot, leaving the grids are at -3V or so. Cathode current is around 180mA idle per tube and as a result the B+ has dropped to 360V. What a disaster! The little trimpots as specified in the DiyTube Eiclone shorted out. I shudder to think of what's been happening to my Sovtek 6L6WXT+'s with that amount of current. I'll probably buy some 2W 25K chassis-mount models and find a safe place to drill a couple holes.
carbon film is fine, it's the value that'll need tweaking. At your 180mA, it's only running slightly over rated dissipation @ 0.32W (if it's 1/4 watt). Bumping that up to 22 or 27 ohms @ 1/4 watt will probly be good enough and will burn up quickly should you lose bias. Don't use 1/2 watt in this case...smaller is better.
good luck
good luck
ray_moth said:
Yeah, unfortunately I can confirm this from experience, and boy they really burn baby..
In particular the ones RS sells are extremely flammable.. 
The aftermath is that they leave quite a mess behind them as a result of their violent demise.There are metal oxide fusible resistors available, however in my experiments they proved to be somewhat non-linear (measured thd) and produced a gritty sound quality when used in the cathode circuit of one of my amplifiers..
Frankly a good old fashioned fuse may sound better.. You can fuse it well beyond the normal operating current to avoid nuisance blowing, but still protect the amplifier when an output tube fails. The safest place for a fuse is usually in the cathode circuit because of the normally low voltages encountered.
sorenj07 said:
I would think that would be fine if on a per tube basis. I have found it helpful sonically to put a small film cap across the fuse (well it seemed to help, nothing scientific about this suggestion therefore YMMV - try with/without if curious) - should have a voltage rating higher than the raw supply even though it's in the cathode circuit, and anything around 0.22uF - 0.47uF should be fine.
Until I get these pots and a drill to mount them on my chassis, I cut the 100K bias connection resistors and brought them to ground. I cut out the little biasing boards and brought each 6L6's cathode to ground with a 470 ohm 5W resistor bypassed by a 470uF 63V electrolytic. Took out the 5U4 tubes and put two 1N4007's in series for each leg of the HV winding. Not my intended "final" fix but for now it's working OK. There's some hum but I think it's from ground loops in my crappy dorm wiring. For example, connecting my powered USB hub to my macbook makes hum increase a TON.
First thing - reconnect the chassis/bus ground to earth. I'd cut it but measured 10-15VAC between the two, and touching the amps increased hum a bit. This is my left monoblock by the way, the one that had buzz on top of the hum and noise. I've since installed RCA brown-base 5R4GYB's.
Then I measured plate voltage on all the tubes and found a stupid wiring mistake, where the 47K plate resistors in the driver stage weren't in parallel - I had 47K going to one plate, then 47K off that PLATE to the plate of the other tube - in effect, one tube had 94K. I'm surprised the amp worked at all the way it was set up. I fixed this but didn't notice much difference in buzz.
Then, another "doh!" moment. It turns out, that I stupidly had left the negative bias tap off the B+ connected, thinking that with the tubes disconnected, no current would be drawn. Of course, there's a bleeder resistor after the rectifier to ground. One of the capacitors was only rated for 160V and was seeing 300V on it, and was bulging. I disconnected the coupling cap from B+ and let that wreckage sit on the ground bus - can't do any harm the way it is. I had hoped that this would have fixed my problem, but I think it only decreased buzz a bit.
I'd been using a portable CD player to test the amp but when I connected it back to my laptop, I had the most horrible hashy noise on top of the buzz. I tried cutting earth ground, which made my laptop have only as much buzz as with the CD player, but had 15VAC on the chassis again!
Damn, this ain't very easy. The only nice news is that with a B+ of 420V, and a 30V drop on my 470 ohm bias resistors for the KT77's, I'm dissipating a comfortable 25W (64mA per). I'm considering just keeping cathode bias because this lower plate voltage/higher current seems a bit closer to class A - am I right?
Anyway I have work to do but I'll keep plugging away at this amp (the left one) until I get it shipshape. I wish I had my workshop - I'm soldering on the floor in my dorm hallway!
Then I measured plate voltage on all the tubes and found a stupid wiring mistake, where the 47K plate resistors in the driver stage weren't in parallel - I had 47K going to one plate, then 47K off that PLATE to the plate of the other tube - in effect, one tube had 94K. I'm surprised the amp worked at all the way it was set up. I fixed this but didn't notice much difference in buzz.
Then, another "doh!" moment. It turns out, that I stupidly had left the negative bias tap off the B+ connected, thinking that with the tubes disconnected, no current would be drawn. Of course, there's a bleeder resistor after the rectifier to ground. One of the capacitors was only rated for 160V and was seeing 300V on it, and was bulging. I disconnected the coupling cap from B+ and let that wreckage sit on the ground bus - can't do any harm the way it is. I had hoped that this would have fixed my problem, but I think it only decreased buzz a bit.
I'd been using a portable CD player to test the amp but when I connected it back to my laptop, I had the most horrible hashy noise on top of the buzz. I tried cutting earth ground, which made my laptop have only as much buzz as with the CD player, but had 15VAC on the chassis again!
Damn, this ain't very easy. The only nice news is that with a B+ of 420V, and a 30V drop on my 470 ohm bias resistors for the KT77's, I'm dissipating a comfortable 25W (64mA per). I'm considering just keeping cathode bias because this lower plate voltage/higher current seems a bit closer to class A - am I right?
Anyway I have work to do but I'll keep plugging away at this amp (the left one) until I get it shipshape. I wish I had my workshop - I'm soldering on the floor in my dorm hallway!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.