Just finished this amp, which was a commission by a friend's family. Push-pull 6P14P's driven by a 6N1P in LTP with a CCS. Sounds very very clear and neutral, a lot more neutral than the amp I built for my dad (but don't tell him that :thmbsp: ) It mates nicely with my Cerwin-Vega D7's, I think.
I wanted to keep things compact but have as solid of a circuit as possible. It's a 425V power transformer choke input rectified by a 5AR4 into a 193J, then 500uF of capacitance. Unfortunately, no smaller chokes for the LTP, but the 82K plate resistors should take care of most ripple. 12x8x2" chassis. The top is actually buckling a bit under the weight of the transformers. The OPT's are Edcor CXPP25-8-7.6K's. Great deals at $50 each including shipping.
Everything is point-to-point. I even "fly" the bias pot and just "tree-root" everything off of it. At least it's accessible
I was going crazy thinking that my mediocre wiring job on the heaters was shorting to chassis. I spent at least 20 minutes pulling, pushing, poking, etc. until I realized that the center-tap of the filament transformer connects to ground
Some pics. I can post the schematic later, it's kind of messy right now.
I wanted to keep things compact but have as solid of a circuit as possible. It's a 425V power transformer choke input rectified by a 5AR4 into a 193J, then 500uF of capacitance. Unfortunately, no smaller chokes for the LTP, but the 82K plate resistors should take care of most ripple. 12x8x2" chassis. The top is actually buckling a bit under the weight of the transformers. The OPT's are Edcor CXPP25-8-7.6K's. Great deals at $50 each including shipping.
Everything is point-to-point. I even "fly" the bias pot and just "tree-root" everything off of it. At least it's accessible
I was going crazy thinking that my mediocre wiring job on the heaters was shorting to chassis. I spent at least 20 minutes pulling, pushing, poking, etc. until I realized that the center-tap of the filament transformer connects to ground Some pics. I can post the schematic later, it's kind of messy right now.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
\An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Sucky news. The rectifier tube has been arcing.
It first happened when I had run the amp loud for an hour or two. A friend brought a different brand 6N1P to try substituting in the driver circuit. I shut off the main power, but only waited for a few seconds before turning the thing back on with the new 6N1P in one channel. The Sovtek 5AR4 arced all over the place (it looked like it was UNDER the tube as opposed to inside it though). I noticed that one of the output tube's cathode resistor and bypass caps had come unsoldered from the ground post. I resoldered it and checked all other connections but the thing has been arcing steadily ever since.
Here's what I do know. I just tried a JJ and it arced again, after the tube had warmed up a bit. Before the JJ arced, I'd measured all transformer windings and found no shorts. However, before the arcing, I'd seen some glowing wires in the 6P14P's which makes me suspect that I'm running them too hot, and blowing the rectifier like that. B+ is around 380V, cathode resistors are 182 ohms plus half a 100 ohm trimpot. According to my EL84 UL sheet I'm not over dissipation... My first impulse here is to add 47-100 ohms from each cathode to ground, after checking all resistors for open connections or shorts, whichever might cause arcing. I have one week before I leave the country for a couple of months so this is kind of urgent...
It first happened when I had run the amp loud for an hour or two. A friend brought a different brand 6N1P to try substituting in the driver circuit. I shut off the main power, but only waited for a few seconds before turning the thing back on with the new 6N1P in one channel. The Sovtek 5AR4 arced all over the place (it looked like it was UNDER the tube as opposed to inside it though). I noticed that one of the output tube's cathode resistor and bypass caps had come unsoldered from the ground post. I resoldered it and checked all other connections but the thing has been arcing steadily ever since.
Here's what I do know. I just tried a JJ and it arced again, after the tube had warmed up a bit. Before the JJ arced, I'd measured all transformer windings and found no shorts. However, before the arcing, I'd seen some glowing wires in the 6P14P's which makes me suspect that I'm running them too hot, and blowing the rectifier like that. B+ is around 380V, cathode resistors are 182 ohms plus half a 100 ohm trimpot. According to my EL84 UL sheet I'm not over dissipation... My first impulse here is to add 47-100 ohms from each cathode to ground, after checking all resistors for open connections or shorts, whichever might cause arcing. I have one week before I leave the country for a couple of months so this is kind of urgent...
An externally hosted image should be here but it was not working when we last tested it.
I did a quick PSUD2 sim, guessing at some of the resistances, and it showed voltage across the rectifier to be pretty high (1200V) and the supply fairly unstable to a current step. You might have some oscillations going on, which start the supply bouncing arond and killing the tube. Or the tube might have had some birth defects and committed suicide.
Have you stuck a scope on things to see what's going on with the output and the rails?
Have you stuck a scope on things to see what's going on with the output and the rails?
The thing is, what mystifies me is that the amp ran fine for a week or so before this happened. Is it possible that one of the LM334's blew? This wouldn't cause arcing would it?SY said:
I do have a scope but the thing is, where should I stick the probes? I haven't picked up any dummy loads yet..
I tried one more time with the blown Sovtek. I'd added 220 ohms to each 6P14P's cathode to treat the PSU very nicely. I also added a CL-90 into the hot side of the PT's primary.
This is what happened. The amp started up nice and slow, with heaters slowly warming up. I turned on the CD player and pressed play, then the 5AR4 arced again. I think it's the LM334. What do you think? My DMM is out of batteries so I'll have to wait to tomorrow to test cathode resistances in the LTP.
This is what happened. The amp started up nice and slow, with heaters slowly warming up. I turned on the CD player and pressed play, then the 5AR4 arced again. I think it's the LM334. What do you think? My DMM is out of batteries so I'll have to wait to tomorrow to test cathode resistances in the LTP.
Update. It's working again! I figured that a bit over 400 ohms per cathode ought to keep dissipation comfortable when B+ is as high as it is. I changed the first capacitor in the choke input to 47uF, and used a 100 ohm 10 watt resistor to connect to the 500uF can cap. A tiny arc when turning on but otherwise, no problems. If the tube arcs again sometime during the next week, I'll bypass the can cap and use 100uF or so instead.
sorenj07 said:
Assuming the four triodes take 20mA, the four penthodes 200mA, Rdc of the transformer is about 100 ohms, Rdc of the 10H is 82 ohm, you could try this setup for better regulation and less stress on the rectifier.
Put a 3uF/630V (oil) cap between the 5AR4 and the choke ;
replace the 500uF with a 100uF ;
Ripple on the 380V B+ is about 1.0Vtt;
ripple on the 344V B++ is about 0.02Vtt;
Stability should be investigated when all currents are known.
Not that.
If the 334 blows, the worst-case current is 2x B+/82k. Assuming something like 350V for B+, that's only 8-9mA. EC8010 has suggested a good idea and it will probably fix the problem, but the basic problem is that you're running a rectifier on its edge with a fundamentally unstable supply. Snubbing will help, as will dropping the 500u cap to a more sensible value (ripple isn't super-critical for a p-p design like this). Don't worry about fancy oil caps or whatever on the input- use high voltage polypropylenes or motor-run, whatever's convenient.
sorenj07 said:
Tuning the choke improves its effect. See: http://www.diyparadise.com/tunedchoke.html
In my previous answer I suggested the use of a higher voltage capacitor which I named 'oil type'. The rated voltage is more important than the dielectricum because there's real highish voltage on it.
Using some guesses for resistances and running a PSUD2 sim, it looks like a 0.47u input cap is too small and the instability is still there. 3u really makes the voltage soar (450V) but stabilizes things. 1u gets you about 380V and a stable supply in the Sim World.
In the real world, you'll have to experiment with the cap size to get the tradeoff right; you'll want to bounce the current deamnd on the rail a bit and see what the response looks like.
At the same time, the inductive kickback is not your friend, especially with 1200V across that poor rectifier tube. You'll have to snub that.
L-input supplies have their virtues but they're not trivial to design and implement.
In the real world, you'll have to experiment with the cap size to get the tradeoff right; you'll want to bounce the current deamnd on the rail a bit and see what the response looks like.
At the same time, the inductive kickback is not your friend, especially with 1200V across that poor rectifier tube. You'll have to snub that.
L-input supplies have their virtues but they're not trivial to design and implement.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.