Re: Question about tube dissipation
Maybe. I have not been able to run my 6L6GC, 6CA7EH, or 6p3s-e much over 20 watts or so without seeing a little red plate.
You should do your calculations using the voltage across the tube (plate voltage minus cathode voltage) and total current through the tube.
That is a good way to calculate total tube current.
More or less. Keep in mind that as the current draw increases, the power supply will "sag" and plate voltage goes down a little.
Voltage drop across the output transformer depends on its DC resistance, not its 5k AC impedance. I'll assume you actually measured the 430 volts at the plate, and that is not a calculated value.
If you have 430 plate volts and 32 cathode volts, then your total dissipation is (430-32)*0.057 = 22.7 watts.
I wouldn't. As I said, all the tubes I've tried in the 30 watt class seem to begin red plate once I get much above 20 watts total. Maybe I've just got crummy tubes. The real test is to let the amp warm up a little, then turn off all the lights in the room. Carefully inspect the plates for any hints of red plate glow. Beam tubes will start to glow first near the seams where the two halves of the plate structure are joined together.
The closer you get to maximum dissipation, the shorter the tube's lifespan. If you are at red plate, the temperature is to the point where outgassing is probably occurring. I'm pretty sure this will poison a tube fairly quickly, and the lifespan will likely be measured in hundreds of hours instead of thousands.
ChrisMmm said:
Maybe. I have not been able to run my 6L6GC, 6CA7EH, or 6p3s-e much over 20 watts or so without seeing a little red plate.
You should do your calculations using the voltage across the tube (plate voltage minus cathode voltage) and total current through the tube.
That is a good way to calculate total tube current.
More or less. Keep in mind that as the current draw increases, the power supply will "sag" and plate voltage goes down a little.
Voltage drop across the output transformer depends on its DC resistance, not its 5k AC impedance. I'll assume you actually measured the 430 volts at the plate, and that is not a calculated value.
If you have 430 plate volts and 32 cathode volts, then your total dissipation is (430-32)*0.057 = 22.7 watts.
I wouldn't. As I said, all the tubes I've tried in the 30 watt class seem to begin red plate once I get much above 20 watts total. Maybe I've just got crummy tubes. The real test is to let the amp warm up a little, then turn off all the lights in the room. Carefully inspect the plates for any hints of red plate glow. Beam tubes will start to glow first near the seams where the two halves of the plate structure are joined together.
The closer you get to maximum dissipation, the shorter the tube's lifespan. If you are at red plate, the temperature is to the point where outgassing is probably occurring. I'm pretty sure this will poison a tube fairly quickly, and the lifespan will likely be measured in hundreds of hours instead of thousands.
Thanks Ty, thats clarified things for me.
I will re-measure the voltages sometime as I have changed tubes since original measurements tho both sets are EL34. My current tubes are Groove Tube EL34M which I believe should be fairly robust. I may try lower value resistors and check each time for "red spot".
Understand what you are saying about tube life but I also got the impression that higher dissipation gave better performance, or possibly slightly greater power delivery for same distortion. Actually very similar to the Pass Mini Alepf I recently built - you aim for 25-30 watts dissipation per Mosfet for optimal performance, trade off here is heatsink size!
BTW, thanks for puting up your SimpleSE - all the pics were very helpful in getting the best layout.
Cheers
I will re-measure the voltages sometime as I have changed tubes since original measurements tho both sets are EL34. My current tubes are Groove Tube EL34M which I believe should be fairly robust. I may try lower value resistors and check each time for "red spot".
Understand what you are saying about tube life but I also got the impression that higher dissipation gave better performance, or possibly slightly greater power delivery for same distortion. Actually very similar to the Pass Mini Alepf I recently built - you aim for 25-30 watts dissipation per Mosfet for optimal performance, trade off here is heatsink size!
BTW, thanks for puting up your SimpleSE - all the pics were very helpful in getting the best layout.
Cheers
Playing with cathode resistor
Out of interest I remeasured voltages, changed the cathode resistor and re-measured, here are the results:
560ohm cathode resistor:
B+ 445V
Plate 425V
Cathode 31.5V
Dissipation 22.1Watt
510ohm cathode resistor:
B+ 440V
Plate 417V
Cathode 30.5V
Dissipation 23.1Watt
I was surprised how much the B+ and plate voltages dropped tho, I recall now, this was just what happened when I was playing around with my Mini Aleph. Also observed tubes and saw no red spot tho I am not entirely sure what this would look like - would I see some red glow in patches?
While I changed the resistors I took the opportunity to put some 2 pin connectors on the board so I can change the resistors easily. Simply done, just drilled an extra hole in the board next to the resistor holes for the 2nd pin. Where the extra hole was on the earth trace I scraped away some of the mask to solder the 2nd pin. On the other pad I soldered a piece of wire between the pins - that way I can use either hole in the connectors and allow for different size resistors if I need to.
This exercise was partly preparation for the arrival of the re-issue Genalex KT88s I ordered. These will have a much greater current draw than the EL34 which in turn will mean the B+ is going to take a dive. I may compensate by lowering the resistance of the resistor in my CLCRC supply and keep it up around the 450v level. Alternatively I can use a different tap on the power supply trafo primary.
Out of interest I remeasured voltages, changed the cathode resistor and re-measured, here are the results:
560ohm cathode resistor:
B+ 445V
Plate 425V
Cathode 31.5V
Dissipation 22.1Watt
510ohm cathode resistor:
B+ 440V
Plate 417V
Cathode 30.5V
Dissipation 23.1Watt
I was surprised how much the B+ and plate voltages dropped tho, I recall now, this was just what happened when I was playing around with my Mini Aleph. Also observed tubes and saw no red spot tho I am not entirely sure what this would look like - would I see some red glow in patches?
While I changed the resistors I took the opportunity to put some 2 pin connectors on the board so I can change the resistors easily. Simply done, just drilled an extra hole in the board next to the resistor holes for the 2nd pin. Where the extra hole was on the earth trace I scraped away some of the mask to solder the 2nd pin. On the other pad I soldered a piece of wire between the pins - that way I can use either hole in the connectors and allow for different size resistors if I need to.
This exercise was partly preparation for the arrival of the re-issue Genalex KT88s I ordered. These will have a much greater current draw than the EL34 which in turn will mean the B+ is going to take a dive. I may compensate by lowering the resistance of the resistor in my CLCRC supply and keep it up around the 450v level. Alternatively I can use a different tap on the power supply trafo primary.
Re: Playing with cathode resistor
If I recall correctly, you used the Hammond 374BX power transformer. I used the same. I estimated the whole power supply has about 730 ohms DC resistance on my amp. If you draw an extra 7.1 mA (both output tubes combined), that equates to an extra 5 volt drop on the B+ rail. It seems your measurements are spot on.
Red plates look something like the tube on the left. Click on the photo for the full sized version.
Yes, the voltage will drop a lot with the KT88. With 560 ohm cathode resistors, I think I had about 71 mA per tube, and probably lost an extra 15 volts off the B+ (compared to the 6CA7 tubes I had before).
ChrisMmm said:
If I recall correctly, you used the Hammond 374BX power transformer. I used the same. I estimated the whole power supply has about 730 ohms DC resistance on my amp. If you draw an extra 7.1 mA (both output tubes combined), that equates to an extra 5 volt drop on the B+ rail. It seems your measurements are spot on.
Red plates look something like the tube on the left. Click on the photo for the full sized version.
Yes, the voltage will drop a lot with the KT88. With 560 ohm cathode resistors, I think I had about 71 mA per tube, and probably lost an extra 15 volts off the B+ (compared to the 6CA7 tubes I had before).
I have a switch on the front to select cathode resistor value. Works a treat! See here: Simple SE switcheable cathode resistor
ChrisMmm said:
Blue glow is usually not a problem, unless it's a specific shade of blue. The deep rich royal blue is OK. I see it all the time in many of my tubes, from the small signal 12A_7 guys up to the big KT88. I believe it is caused by electrons that got accelerated so fast they "missed" the plate and hit the glass instead. It causes something in the glass (impurities? minerals?) to fluoresce with a distinctive shade of blue. I can say with much certainty that if you have that kind of blue glow (and not all tubes do!) then the effect is more pronounced at higher voltages.
Here's a tube with a happy blue glow near the mica spacers.
I've also noticed that some tubes show light blue stripes on the inside of the plate where the electron beam hits. This seems to happen most with the Russian tubes. It has something to do with the composition of the metal used in the plates, and does not indicate any problems. My photo is too blurry to show it, but the light blue color seen through the plate holes is actually a series of stripes corresponding to the pitch of the grids. Maybe if you look carefully, you can see the stripes in the rightmost 6p14p in this photo:
There is a kind of glow that is a problem, but I think it tends to be more of a pinkish/whitish-blue color. If you see that shade floating in the space between the grid and the plate, it might indicate gas in the tube. That tube is done, and doesn't have any life left to give. A few people have suggested that baking such a gassy tube in the oven might allow the getter to soak up the gas, but I suspect the chances are low.
Edit... here's another photo of my Simple SE with a pair of 6CA7EH. The tube on the left shows a nice blue glow around the plate holes, while the tube on the right shows nothing. This goes back to the whole "some tubes do it, some don't" thing.
my russian 6p14ps have exactly that 'inner plate' glow. bright blue. doesnt seem to do any harm at all.
Couple of things changed on my SSE.
First, having checked with Hammond it was OK, I rewired the power trafo primary to match the local 230v mains voltage (original wiring instructions were either 220 or 240v, I had 240v). This gives me some 18v extra on the B+. So with EL34M and 510ohm cathode resistor I get:
B+ 558V
63ma current draw
25 watt dissipation
No problem with red spot. Listened to this config for an hour or so. Big powerful sound, what impressed me over and above what my other single ended amps sound like is the dynamics at lower frequencies. I am listening at near concert levels with my Fostex BIB speakers - this is awesome!
Next my re-issue Genelex KT88s arrived. Installed with the existing 510ohm resistors and get:
B+ 447V
76.5ma current draw
29 watt dissipation
Tried a 390ohm resistor briefly:
B+ 436V
90ma current draw
33 watt dissipation
I went back to the 510ohm resistor for some listening. First the KT88s seemed to require a notch or 2 more on the volume control - could this combination result in less overall gain? I wouldn't say this was better than the EL34M after the brief listening session but the incredible dynamics are still there. More listening to come. I have read these tubes need some burn-in, any experience with this?
With the KT88s and 510ohm cathode resistor the power trafo is getting fairly hot, probably around 55C. Now I know from experience this happens with class A amps but this is quite a bit hotter than my EL84 amp tho that has less current draw and greater current capacity trafo. I am not too concerned but just wondering if I should be OK running for a longer period of time like that (say 6 hours). I would hope that should be OK??
First, having checked with Hammond it was OK, I rewired the power trafo primary to match the local 230v mains voltage (original wiring instructions were either 220 or 240v, I had 240v). This gives me some 18v extra on the B+. So with EL34M and 510ohm cathode resistor I get:
B+ 558V
63ma current draw
25 watt dissipation
No problem with red spot. Listened to this config for an hour or so. Big powerful sound, what impressed me over and above what my other single ended amps sound like is the dynamics at lower frequencies. I am listening at near concert levels with my Fostex BIB speakers - this is awesome!
Next my re-issue Genelex KT88s arrived. Installed with the existing 510ohm resistors and get:
B+ 447V
76.5ma current draw
29 watt dissipation
Tried a 390ohm resistor briefly:
B+ 436V
90ma current draw
33 watt dissipation
I went back to the 510ohm resistor for some listening. First the KT88s seemed to require a notch or 2 more on the volume control - could this combination result in less overall gain? I wouldn't say this was better than the EL34M after the brief listening session but the incredible dynamics are still there. More listening to come. I have read these tubes need some burn-in, any experience with this?
With the KT88s and 510ohm cathode resistor the power trafo is getting fairly hot, probably around 55C. Now I know from experience this happens with class A amps but this is quite a bit hotter than my EL84 amp tho that has less current draw and greater current capacity trafo. I am not too concerned but just wondering if I should be OK running for a longer period of time like that (say 6 hours). I would hope that should be OK??
ChrisMmm said:
I'm using the same Hammond 374BX power transformer. My line voltage is about 125 VAC, 60 Hz. I have mine wired for 120 volt operation. With KT88 tubes and 560 ohm cathode resistors, I see about 72 mA current through the tube. I don't recall the B+. After several hours the power transformer gets very hot. You can keep your hand on it, but you'd really rather not.
At 90 mA per output tube, plus 10 mA per side on the driver tube, you're pulling 200 mA out of the transformer. The 374BX is rated for 175 mA. I don't know whether these transformers are rated conservatively. I'd suspect not. If it were my amp, I'd be inclined to stay with the 510 ohm resistors just to keep the current draw down a little bit.
Is your AC line at 50 Hz? I think that will make the transformer run even hotter, although the 374BX is supposedly rated for 50 Hz operation.
Another thought occurred to me - I believe Gery rates the Transcendar 10 watt 5K:8 output transformer for 90 mA DC on the primary. I think if you exceed that you might start to see the effects of saturation on the core sooner than you'd like.
My Hammond 374BX gets very hot with the KT88s too. Runs cooler with the E34Ls and 6L6s. Interested in your final listening results. I thought that the mid range detail was better with the E34Ls, bass is handled by SS amps in my system.
Cheers,
Chris
Cheers,
Chris
About the Hammond 374BX, If I use another transfo - Hammond 266p6A - for the filament only (KT88), do you think the 374BX will run cooler ?
Tigrine
Excuse me for my english, my french is better!
Tigrine
Excuse me for my english, my french is better!
Ty_Bower said:
Ty, isn't that 175ma accross the whole 750v winding which wil be 350ma at 375v? Its a 200VA transformer so 15va for the 5v winding, 25va for the 6.3v. The the main winding is 750v x 175ma is 130va - total 170va as I make it.
Yes, 50hz so its is derated a little.
I only tried the 390ohm to see what the voltages looked like, I will almost certainly stick with 510 or similar. I could go the "turbo switch" way but I am not keen on the extra switch and wiring.
Oh, I came accross your earlier post saying the KT88 needed a little more on the volume control so that puts my mind at peace - I am not going crazy or made a silly mistake (well not in this case). BTW I am using a autoformer volume control with no gain and its about as loud as I want to listen at 2 or 3 o'clock on the volume knob - so the overall gain structure of my system is just about where I want it.
Out of interest do you understand why the gain in the amp has changed with the KT88?
chrish said:
Like I say I wasn't immediately blown away with the KT88s but a lot more listening is needed to decide. I will be dissapointed if they are not better than the Groove Tube EL34M as the KT88s are twice the price. But, hey, ya gotta try these things!
tigrine said:
Marginally cooler I would expect, you are only droping 20va load off it and ita rated 200va. I don't think it would be worth the effort.
I would guess that the gain is different because the two valves have different characteristics. There may be other considerations, but the gain of the EL34 is 11, while the KT88 is 8.
Chris
Chris
chrish said:
Thanks, I was trying to find that out - wher did you get the gain figures from? I looked in spec sheets but didn't see it.
I thought voltage gain was in the driver section and the power tube was for current gain or have i got that all wrong?
Frank's Electron Tube Data Sheet Page
Has downloadable tube data sheets.
You would have to plot the load lines of each tube against the driver input to see exactly what is happening.
Has downloadable tube data sheets.
You would have to plot the load lines of each tube against the driver input to see exactly what is happening.
Another simple (wink, wink) option would be to drop a different rectifier in like a 5U4 that would drop more voltage to reduce the current draw. Granted, the 5U4 doesn't have the lovely slow rise time of the 5AR4 but life is a series of compromises.
Hey, not such a bad idea! My listening habits are lower level background listening in the weekends for long periods with serious listening at high level week night evenings. So the 5AR4 for week nights and a 5U4 for weekends, works well!
And an excuse to buy another tube!!
Another possibility may be switching different resistors in the CLCRC PS filter. That would require minimal extra wiring in a less critical area.
And an excuse to buy another tube!!
Another possibility may be switching different resistors in the CLCRC PS filter. That would require minimal extra wiring in a less critical area.
- Status
- Not open for further replies.