Hey guys,
I just got my RA-100 amp fired up (Thanks Jim!),but there's a small problem..
It keeps vaporizing cathode resistors!
For testing,I'm using Electro Harmonix EL34's. I assumed they'd be up to dealing with 600V B+,but maybe I'm wrong.
This amp was originally designed for 6L6GC's.
At first, I used some 2W "fusible" metal-film cathode resistors..(they're about the size of regular 1/2W R's.. 😕 ) one of those Nuked,so I replaced it with the last 10ohm R I had on hand..And it nuked another resistor...So then I went to Radioshack,and found some 1W Metal-oxide resistors.. Then both of those nuked in one channel.
So then I had 2x 2W's in one channel and 2x 1W's in the other...and things seemed fine for a while...until another of the 1Watters went *POP!*
Pondering over this for a while,My theory is that the EL34's are arcing from the plate to G3 (I have G3 and cathode (pins 1/8) tied together,might that be an issue?),thus dumping the energy from the filter caps into the poor cathode R's. and violently exploding them -That's the only thing that I can think of that would cause such a loud *POP!* to happen.. It's fairly loud,and sometimes sparks fly! (scared the crap outta me a couple times!!)
I don't think it is tube related,because it's not on the same socket/tube all the time,it seems totally random.
This usually happens when I flip the standby switch,but not always. The last time a resistor nuked,it had been on for a couple of minutes,I had just started playing some music through it (low volume)..and a minute later *POW!* -another resistor gone!
That's 5 so far!
The bias/grid voltage on the output tubes checks out.. Supposed to be ~ -72V,for 50ma per tube(6L6GC).. IIRC it's set somewhere ~ -76V right now,with about 25ma per tube(using EH EL34's)...so it's even under-biased a bit.
I've even put the 'scope across the cathode R's,and watched the transient at 'turn-on'.. not much,a quick jump to ~500mv,then it drops right back to ~250mv.
I have NOT 'scoped for oscillations or any parasitic stuff yet..
Could suddenly bursting into oscillation completely nuke the cathode R's like that? (I may add grid stoppers,etc. anyways.)
Any other thoughts on what might be going on here?
Has anyone else run EH EL34's at 600V and had any problems?
Does anyone have any suggestions for 6L6GC's that will stand up to 600V?
(I've even thought about switching to 807's,but I'm not sure the G2 would be up to it,since this amp is ultra-linear,with ~600V on G2's!)
For those who are interested,
Here's the Original Design by Fred Nachbaur.
Here are some Pictures,etc. of my amp.
Also,a thread on Geek's forum about it's (slow) progress..
Thanks! Patrick.
Edit:
Should also mention that I tried tapping the tubes and chassis,thinking that maybe a grid wire was loose.. nada..
also wiggled them in the sockets a bit,thinking maybe a loose contact,also nada -the sockets are quite a tight fit,actually.
I just got my RA-100 amp fired up (Thanks Jim!),but there's a small problem..
It keeps vaporizing cathode resistors!
For testing,I'm using Electro Harmonix EL34's. I assumed they'd be up to dealing with 600V B+,but maybe I'm wrong.
This amp was originally designed for 6L6GC's.
At first, I used some 2W "fusible" metal-film cathode resistors..(they're about the size of regular 1/2W R's.. 😕 ) one of those Nuked,so I replaced it with the last 10ohm R I had on hand..And it nuked another resistor...So then I went to Radioshack,and found some 1W Metal-oxide resistors.. Then both of those nuked in one channel.
So then I had 2x 2W's in one channel and 2x 1W's in the other...and things seemed fine for a while...until another of the 1Watters went *POP!*
Pondering over this for a while,My theory is that the EL34's are arcing from the plate to G3 (I have G3 and cathode (pins 1/8) tied together,might that be an issue?),thus dumping the energy from the filter caps into the poor cathode R's. and violently exploding them -That's the only thing that I can think of that would cause such a loud *POP!* to happen.. It's fairly loud,and sometimes sparks fly!
(scared the crap outta me a couple times!!)I don't think it is tube related,because it's not on the same socket/tube all the time,it seems totally random.
This usually happens when I flip the standby switch,but not always. The last time a resistor nuked,it had been on for a couple of minutes,I had just started playing some music through it (low volume)..and a minute later *POW!* -another resistor gone!
That's 5 so far!The bias/grid voltage on the output tubes checks out.. Supposed to be ~ -72V,for 50ma per tube(6L6GC).. IIRC it's set somewhere ~ -76V right now,with about 25ma per tube(using EH EL34's)...so it's even under-biased a bit.
I've even put the 'scope across the cathode R's,and watched the transient at 'turn-on'.. not much,a quick jump to ~500mv,then it drops right back to ~250mv.
I have NOT 'scoped for oscillations or any parasitic stuff yet..
Could suddenly bursting into oscillation completely nuke the cathode R's like that? (I may add grid stoppers,etc. anyways.)
Any other thoughts on what might be going on here?
Has anyone else run EH EL34's at 600V and had any problems?
Does anyone have any suggestions for 6L6GC's that will stand up to 600V?
(I've even thought about switching to 807's,but I'm not sure the G2 would be up to it,since this amp is ultra-linear,with ~600V on G2's!)
For those who are interested,
Here's the Original Design by Fred Nachbaur.
Here are some Pictures,etc. of my amp.
Also,a thread on Geek's forum about it's (slow) progress..
Thanks! Patrick.
Edit:
Should also mention that I tried tapping the tubes and chassis,thinking that maybe a grid wire was loose.. nada..
also wiggled them in the sockets a bit,thinking maybe a loose contact,also nada -the sockets are quite a tight fit,actually.
DigitalJunkie said:
I'm trying to follow the schematic for the RA-100 power supply. I see the standby switch near the top, slightly to the left of center. I'm immediately suspicious of standby switches, especially if the power supply has anything other than an indirectly heated tube rectifier (like a 5AR4).
The standby switch seems like it's in an inopportune place. Why not try putting it before the rectifying diodes, with a 1M or so resistor and a .1uF 1KV+ cap across it? this might help damp any voltage spikes. This is all second-hand, by the way, just what I've been told - kind of a rule of thumb.
sorenj07 said:
He needs to mind the bias supply. You can't just put a single pole standby switch on the center tap, as it will take out the bias voltage too. You want to keep the bias supply running even when you're on standby. He would have to use a double pole switch and break both sides of the secondary going to the rectifiers.
I'll agree, though. Living on a 600 volt DC rail doesn't seem like a happy place for an ordinary switch. I hope someone doesn't someday forget to turn off the main power, and then flip the switch to standby while the amp is running!
Patrick,
what you need with 600 Volt B+ are different toobs !
The only ones that would stand 600V on the screens are KT88.
I think the screens of your EL34 do arcs since they are rated at 425V only.
Cheers Klaus
what you need with 600 Volt B+ are different toobs !
The only ones that would stand 600V on the screens are KT88.
I think the screens of your EL34 do arcs since they are rated at 425V only.
Cheers Klaus
Uhm,I've already done that a couple times.. Why would that be an issue?! (switch contacts arcing over?) I haven't had any issues with it yet.
I think I might replace the standby switch with a beefier one,just because.
The R/C across the switch contacts might be a good idea too.
I can't really move the switch,due to the bias issue that Ty mentioned.
6550's might be another possible candidate,except I may need to add a second/separate heater transformer,the Hammond 282X is rated for 6A on the 6.3V winding.
I wonder if "fat bottle" EL34's might take the abuse a bit better then the "skinny" ones..
I've heard people say they've run Mullard EL34's at uber-high voltage,and didn't have any problems.... But I dunno if I can afford a quad of Mullards!
Fred apparently used JJ's,and "unknown" chinese 6L6GC's (!) without any problems.
Maybe I'll get a quad of JJ 6L6's,and hope they don't arc/explode.
Jim,George,anyone?, Do you have any suggestions for a robust 6L6GC tube?
The KTxx tubes could be interesting too,I've always wanted to use some in a project,but haven't got around 'tuit yet.
Decisions,decisions..😕
Thanks for the input guys,keep it coming!
600 volts is far above the spec for any flavor of 6L6GC. Don't even think of using an older (6L6GB) based tube (including the 807). They will fry on 500 volts G2 (experience speaks). I have no problem "testing" the limits of cheap tubes.
My biggest power supply only goes to 550 volts, so I have limited experience above this voltage. I have used several of those "unknown" chinese 6L6GC's, also without failure at 550 volts and at dissipations far in excess of the specs. The good ones come from the Shuguang factory and are often labeled Sino. The older "Coke Bottle" shaped ones were the best from an abuse standpoint. These are getting hard to find now. The latest ones that I tested are labeled 6L6GCR. They are in straight sided envelopes that don't handle high dissipations as well as the Coke Bottle version.
I have torture tested several of these without any failures, but I am usually doing this in a situation where I don't care if they blow up.
I have also used the Electro Harmonix KT-88 in a SE amp at 525 volts without issue. I run them at dissipations approaching 40 watts without glow.
FWIW, according to Mullard's data sheet on the EL34, the 'Design Centre' maximum rating in triode mode for the plate (which also means g2) is 600v. In native pentode mode, the g2 rating is only 500v. Those are absolute limits and it would be most unwise, IMHO, to exceed them; indeed, it would be best to stay well below them.
I don't see how the screen voltage is involved with what's happening to your cathode resistors. To blow up a 10 ohm 2w resistor, there would need to be significanly more than 2W dissipated in it. We know that the power dissipated will be current squared multiplied by the resistance, so the current = sqare root of the power divided by the resistance. For 2 watts of power, the current through a 10 ohm resistor = SQRT(2/10) - 0.447 amps. That is a huge current and I don't see how exceeding g2 voltage could give rise to it.
On the other hand, if the control grid bias goes wrong, this could easily cause a big rise in current. In that RA-100 circuit, the bias voltage depends on a very complicated arrangement and the grids are not properly protected against accidents. The two series diodes from grid to ground ensure that the grid will never get more than 1.4v positive if the LTP and CCS ever fail, but the damage will have been done by then.
I don't see how the screen voltage is involved with what's happening to your cathode resistors. To blow up a 10 ohm 2w resistor, there would need to be significanly more than 2W dissipated in it. We know that the power dissipated will be current squared multiplied by the resistance, so the current = sqare root of the power divided by the resistance. For 2 watts of power, the current through a 10 ohm resistor = SQRT(2/10) - 0.447 amps. That is a huge current and I don't see how exceeding g2 voltage could give rise to it.
On the other hand, if the control grid bias goes wrong, this could easily cause a big rise in current. In that RA-100 circuit, the bias voltage depends on a very complicated arrangement and the grids are not properly protected against accidents. The two series diodes from grid to ground ensure that the grid will never get more than 1.4v positive if the LTP and CCS ever fail, but the damage will have been done by then.
Hi Ray,
Exactly. I don't think G2 is related,except that I'll need to use tubes that can handle the ~600V on the Screen grid. -So far KT-88's are looking like the best bet,for UL operation.(and even those would be on the hairy edge of Vg2 spec's.)
I think what's happening is that the Plate (or perhaps G2?) is arcing to G3 (which is tied to the cathode) and frying the daylights out of the cathode resistors.
Basically arcing,and shorting B+ to ground through the cathode R's.
The PSU in this amp has plenty of filtering capacitance,and lots of current available.More than enough to blast a 2W resistor to the moon.😉
One other thought I had,was to regulate the B+ to maybe 500V.
I'm not terribly sure if there is enough room to shoe-horn another regulator in there,but I could try. 😀
I'm also considering going pentode,sort of as a 'last resort'.Might make things a bit easier,and I've already got a regulated +230V supply I could possibly use for the screens. Hmm.
Again,Thanks! -I'm off to look at some more datasheets.
DigitalJunkie said:
You could test this theory by connecting g3 to the plate. Some people will tell you that this is how it should be connected anyway.
That's for triode mode, not UL mode. In any case it's not really relevant here, because you would never connect the suppressor to both plate and cathode.
What Patrick seems concerned about is that, with the suppressor already connected externally to the cathode (at the tube socket), arcing might be taking place internally, between the suppressor and the plate or screen, thus providing a path from the OPT primary to the cathode resistor.
I doubt if that's happening, though - I still suspect the grid bias is going wrong. To my (perhaps over-cautious) mind, that RA-100 circuit looks like an accident waiting to happen.
ray_moth said:
My mistake. For some reason I had the idea that he had the EL34's triode strapped.
EDIT: It just occurred to me that he could still connect g3 to the plate. (Obviously not to both the plate and cathode.) It shouldn't hurt anything. It would simply turn the pentode into a tetrode with the associated kinks in the plate curves.
In any case, I agree that arcing is probably not the problem.
EH EL34s work fine in Music Man guitar amps with 700+ volts on the plate - I doubt plate voltage is an issue.
Screen voltages above 500 can be problematic in some circumstances. But I agree with Ray, I doubt that's the issue as well.
Have you seen any arcing or is there any evidence of arcing in the tubes?
Be careful! With 600 volts from plate to cathode 50 ma will mean 30 watts dissipation. That's 20% over maximum for an EL34.
I'm not familiar with the circuit, but it certainly seems to be a tube killer!
Screen voltages above 500 can be problematic in some circumstances. But I agree with Ray, I doubt that's the issue as well.
Have you seen any arcing or is there any evidence of arcing in the tubes?
Be careful! With 600 volts from plate to cathode 50 ma will mean 30 watts dissipation. That's 20% over maximum for an EL34.
I'm not familiar with the circuit, but it certainly seems to be a tube killer!
To me this is a ****-poor design...sorry if I hurt anyone's feelings...
Even KT88 valves would have a tough time dealing with this...
The screen voltage is simply way too high..... This is way beyond the capabilty of the valves...
A better approach would be to use a voltage doubler arangement to make the 600V then tap off the center of the doubler to put 300V on the screens...
The problem is when the plate is swinging with AC signal....the load-line will swing the plate all the way down to the diode line,thus leaving a large difference between plate to screen voltage... ....The supressor grid is suppose to increase this limitation..but is over whelmed and can't deal such a difference...
Chris
Even KT88 valves would have a tough time dealing with this...
The screen voltage is simply way too high..... This is way beyond the capabilty of the valves...
A better approach would be to use a voltage doubler arangement to make the 600V then tap off the center of the doubler to put 300V on the screens...
The problem is when the plate is swinging with AC signal....the load-line will swing the plate all the way down to the diode line,thus leaving a large difference between plate to screen voltage... ....The supressor grid is suppose to increase this limitation..but is over whelmed and can't deal such a difference...
Chris
I haven't noticed any arcing in the tubes..The flying sparks from the cathode R's captured my attention first.
So,you don't think it's internal arcing in the tubes?
Hmm.. Maybe the bias goes wonky for a second?
(like a loose connection or something. -I wonder if the potentiometers are dirty,or aren't making good wiper contact? They seemed nice and smooth..the grid voltage wasn't 'jumpy' or anything when I was adjusting them,but that might not be saying much.)
Or,Maybe it's bursting into oscillation.
(I installed 1K G1 stoppers,and 470ohm G2 stoppers last night -'just cause'.)
I dunno,I'm kinda stumped.
I might try connecting it up Pentode,and seeing how it behaves then.
Thanks for the feedback guys!
So,you don't think it's internal arcing in the tubes?
Hmm.. Maybe the bias goes wonky for a second?
(like a loose connection or something. -I wonder if the potentiometers are dirty,or aren't making good wiper contact? They seemed nice and smooth..the grid voltage wasn't 'jumpy' or anything when I was adjusting them,but that might not be saying much.)
Or,Maybe it's bursting into oscillation.
(I installed 1K G1 stoppers,and 470ohm G2 stoppers last night -'just cause'.)
I dunno,I'm kinda stumped.
I might try connecting it up Pentode,and seeing how it behaves then.
Thanks for the feedback guys!
Fred has left this earth, an auto accident if I remember correctly.
The arc theory could be tested easily enough. Simply gonnect G3 directly to ground, or to its own 10 ohm resistor to ground. If an arc from G2 pr plate hits G3 it won't take out the cathode resistor. If the resistor on G3 blows there was an arc.
I would think that an arc strong enough to blast a 2 watt resistor would cause permanent damage to the tubes. There should be some missing grid wires. This may not be visible inside the plate, but some evidence may be left floating around inside the tube. Remove each tube and examine them for any loose stuff inside. Look for tiny pieces of wire, or small white flakes loose inside. I am guessing that there is some type of intermittent bias anomaly causing momentary current surges. This could damage the cathodes leading to the loose white flakes inside.
The arc theory could be tested easily enough. Simply gonnect G3 directly to ground, or to its own 10 ohm resistor to ground. If an arc from G2 pr plate hits G3 it won't take out the cathode resistor. If the resistor on G3 blows there was an arc.
I would think that an arc strong enough to blast a 2 watt resistor would cause permanent damage to the tubes. There should be some missing grid wires. This may not be visible inside the plate, but some evidence may be left floating around inside the tube. Remove each tube and examine them for any loose stuff inside. Look for tiny pieces of wire, or small white flakes loose inside. I am guessing that there is some type of intermittent bias anomaly causing momentary current surges. This could damage the cathodes leading to the loose white flakes inside.
- Status
- Not open for further replies.