I recently completely rebuilt my Dyanco ST-70 using a kit from Bob Latino. I brought it up in the usual way...no tubes, check bias/heaters, add rectifier, etc. Each step it was brought up slowly on a variac. Everything seems fine. I did move up to a twist-lock with more capacity.
I was doing something and bumped the power switch on the variac by accident and quickly flipped it back on. When I did this, the rectifier arced for a second or so. Then it was fine. It's a cheap Chinese 5AR4 that I was using for bring-up, but it's brand-new.
It occurred while the cathode was hot, so I am wondering if the current surge was just too much for it. I usually use 5U4 or 5U4GB for tube rectification, which are directly-heated and have never observed this. This is the only amp I have that uses a 5AR4, so I am wondering if the tube is junk or not.
I was doing something and bumped the power switch on the variac by accident and quickly flipped it back on. When I did this, the rectifier arced for a second or so. Then it was fine. It's a cheap Chinese 5AR4 that I was using for bring-up, but it's brand-new.
It occurred while the cathode was hot, so I am wondering if the current surge was just too much for it. I usually use 5U4 or 5U4GB for tube rectification, which are directly-heated and have never observed this. This is the only amp I have that uses a 5AR4, so I am wondering if the tube is junk or not.
Well, you say you increased the capacitance. The more capacitance used, the more prone to arcing. Also, tube rectifiers don't like being flicked off then on, really quick after they have been warmed up. Everything is still conductive because the cathodes are hot. So you get less current limiting.
I'd think that direct heated rectifiers are less prone, but not immune. Their cathodes have less thermal inertia, so they heat quicker and cool quicker.
On a side note, I have heard of poor quality 5ar4's floating around.
I'd think that direct heated rectifiers are less prone, but not immune. Their cathodes have less thermal inertia, so they heat quicker and cool quicker.
On a side note, I have heard of poor quality 5ar4's floating around.
Yeah, that's what I was thinking. After a half-dozen "cold" starts last night, I observed a tiny bit of arcing once. I ran the amp pretty hard for a few hours throughout the evening to do some burn-in and it seemed fine. The tube might have a defect though, because the heater is sticking out of the bottom of the cathode about about 1/4" and doesn't reach the top. The arcing happens down at the bottom, so it might be jumping from the hot heater to the plate.
I need to check the frequency response. It seemed to be lacking in the bass dept, but detail got noticeably better as the evening went on.... It has the VTA driver board and russian PIO coupling caps and so far I am happy with it in that respect.
I need to check the frequency response. It seemed to be lacking in the bass dept, but detail got noticeably better as the evening went on.... It has the VTA driver board and russian PIO coupling caps and so far I am happy with it in that respect.
I have a VTA70 and in the beginning I arced about 4 JJ GZ34s. Had an Audio Sound Labs cap board and the 60uf the rectifier tube was seeing was too much, Lowered it too 41uf and that partially cured the problem.
I always let the amp cool down at least 1/2 hour before a restart, too early there is a surge that has arced my rectifier tube. With a Weber copper top I get 430 volts tops B+ too the board cold start. With a 10 minute quick startup I get 460 volts.
Usually in my experience once a tube arcs it is no good. I now use a sovtek rectifier tube or the Weber and have been trouble free since. I have about 600uf between the cap and VTA boards.
Wait 100 hours before coming to any conclusions about the sound. Sometimes takes at least that long for everything to settle in.
Richard
I always let the amp cool down at least 1/2 hour before a restart, too early there is a surge that has arced my rectifier tube. With a Weber copper top I get 430 volts tops B+ too the board cold start. With a 10 minute quick startup I get 460 volts.
Usually in my experience once a tube arcs it is no good. I now use a sovtek rectifier tube or the Weber and have been trouble free since. I have about 600uf between the cap and VTA boards.
Wait 100 hours before coming to any conclusions about the sound. Sometimes takes at least that long for everything to settle in.
Richard
This is a serious problem with all vintages of 5AR4 - I've seen it in completely stock ST-70 with Mullard or Telefunken 5AR4, it isn't always fatal, but it frequently is particularly in current production types where a cathode to plate short is often the consequence.
Interestingly enough placing a single UF4007 or 1N4007 diode in series with each plate lead will cure this and a lot of other issues with modern rectifier tubes. I do this as a matter of course in everything I build now, and I hardly ever experience a rectifier failure even with current production types. This effectively increases the PIV rating of the rectifier and also makes sure it does not really see much reverse voltage at the plates which helps with the arcing issue. Equalization resistors (normally used to swamp diode reverse leakage currents) are not required at these voltages with the '4007 and a 5AR4 so leave them out.
Interestingly enough placing a single UF4007 or 1N4007 diode in series with each plate lead will cure this and a lot of other issues with modern rectifier tubes. I do this as a matter of course in everything I build now, and I hardly ever experience a rectifier failure even with current production types. This effectively increases the PIV rating of the rectifier and also makes sure it does not really see much reverse voltage at the plates which helps with the arcing issue. Equalization resistors (normally used to swamp diode reverse leakage currents) are not required at these voltages with the '4007 and a 5AR4 so leave them out.
I was doing some tests with a few different tubes. I have some new JJ KT77s, which I have never tried before but had heard they sound good in the Dynaco. I happened to have a JJ 5AR4 as well. I swapped all of the tubes at once, reset the bias and wanted to take them for a spin. Well as soon as the B+ came up the 5AR4 arced-over and the fuse blew. I had the bias meters installed and only one was showing any current. The other 3 were flat-lined. I am not impressed with the build-quality of these KT77s at all. One has a crooked tube and two don't even have a logo printed on them. They all came in a "matched quad", but obviously one or more is shorted. I tried the same rectifier on a set of SED EL34s and they biased-in fine.
Anyway, I ran some tests on it and was very happy with the results. Frequency response at full output (onset of clipping, about 45W into the load):
Distortion at full output:
More plots here:
http://www.knizefamily.net/russ/electronics/electron-tube-audio/dynakit-st70/performance/
Anyway, I ran some tests on it and was very happy with the results. Frequency response at full output (onset of clipping, about 45W into the load):
Distortion at full output:
More plots here:
http://www.knizefamily.net/russ/electronics/electron-tube-audio/dynakit-st70/performance/
kevinkr said:
This topic is illustrated in the following link:
http://www.freewebs.com/valvewizard/fullwave.html
An externally hosted image should be here but it was not working when we last tested it.
I had a similar arcing issue with a 5R4GY at a "hot start". I think it is because of the reverse voltage potential from the input capacitor. I am not familiar with the circuit but if you increased the input capacitance the "bleed off" after the rectifier is turned off may now not be sufficient. When the rectifier is switched back on quickly the voltage of the capacitor is breifly higher than the output of the rectifier causing reverse current? This probably what the arcing is?
With a capacitor input filter the first cap is reccomended to be less than 60uF. The schematic I looked at shows a CLCRCRC filter. The first cap is 30uf Did you increase that to over 60? or even aproaching 60? That might be the issue.
With a capacitor input filter the first cap is reccomended to be less than 60uF. The schematic I looked at shows a CLCRCRC filter. The first cap is 30uf Did you increase that to over 60? or even aproaching 60? That might be the issue.
coldcathode said:
No, not usually. I have had supply circuits with very short time constants arc upon reapplication of power if the filaments were still warm. Interestingly enough this issue can be greatly exacerbated if the output tubes are still hot as their internal resistance is relatively low. I have seen this with choke inputs where there is no capacitor right after the rectifier. I have to assume that this arcing issue relates to minimal clearances in the rectifier tube, powerful electro-static fields in the area between the plate and cathode, and more than likely a flyback pulse in the transformer secondary when power is re-applied that momentarily exceeds the voltage rating of the rectifier - this is even more likely if the power switch has appreciable contact bounce. (Which most do.)
My suspicion is that adding an rc snubber network across the high voltage secondary would not hurt, and as I mentioned before series diodes help a great deal.
Hi Raymond,
I haven't observed any significant switching noise with the ss series diodes, when the vacuum tube rectifier cuts off this disconnects the ss diodes from any large radiating current loop that would otherwise exist. I do recommend UF4007 if this is a concern, and/or a small ceramic cap (2KV min) in the range of 1nF to 10nF (0.01uF) across the ss diodes, but the hexfred is total overkill in this application - might as well eliminate the tube rectifier if going this route. I use mostly 1N4007 for this application and have no noise issues, and no distinguishable difference in performance other than not frying rectifiers on a regular basis.
I haven't observed any significant switching noise with the ss series diodes, when the vacuum tube rectifier cuts off this disconnects the ss diodes from any large radiating current loop that would otherwise exist. I do recommend UF4007 if this is a concern, and/or a small ceramic cap (2KV min) in the range of 1nF to 10nF (0.01uF) across the ss diodes, but the hexfred is total overkill in this application - might as well eliminate the tube rectifier if going this route. I use mostly 1N4007 for this application and have no noise issues, and no distinguishable difference in performance other than not frying rectifiers on a regular basis.
I had this problem with a st70, on advise from a friend I bought a electroharmonix rectifier tube and problem went away. Another crazy story is how I had 80uf on the first stage without any problems
, I used it like this for over a year before I sold it. That electro tube held strong all that time, never changed the tube again.
On a side note I bought the new 6CA7 fat boys from russia, what a sweet change that made over the EL34's I was using. This thread makes me want to source another dynaco clunker
, I used it like this for over a year before I sold it. That electro tube held strong all that time, never changed the tube again.On a side note I bought the new 6CA7 fat boys from russia, what a sweet change that made over the EL34's I was using. This thread makes me want to source another dynaco clunker
Evenharmonics said:
No, oddly enough it isn't and it is really only required with the 5AR4 and strongly recommended when pushing the envelope voltage wise with the 5U4XXX (say 1KV)
I've been doing this for about 10yrs now, and have to admit that Eric Barbour originally suggested this to me when I was working on an Italian made dht based monoblock (572-3) using 5U4 at voltages in excess of 1.2KV. I applied it to the 5AR4 when I was having problems in my own products and this helped greatly despite the fact that I was not running them any harder than is the case in ST-70 and in a lot of cases not half as hard.
I had read on another forum that you may hear "Switching" if you run the two IN4007 Diodes inline from the power supply/Power Transformer to the GZ34 as seen in this thread...
A solution was to use the UF4007 (Fast Recovery) Diodes in place of the 1N4007. this is in regards to running 2 1A / 1000Volt Diodes inline from the Power Transformer to the GZ34 as seen in the diagram of this thread.
I have the IN4007s now... Wondering about doing a switch over to UF4007 as I can do this for a $1.00.
Thoughts??
A solution was to use the UF4007 (Fast Recovery) Diodes in place of the 1N4007. this is in regards to running 2 1A / 1000Volt Diodes inline from the Power Transformer to the GZ34 as seen in the diagram of this thread.
I have the IN4007s now... Wondering about doing a switch over to UF4007 as I can do this for a $1.00.
Thoughts??
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.