Folks,
I'm puzzled by something. Why do most designers here run so wimpy operating points on their 300B's? Just askin'....
I've been designing this 300B amp for a while. Most people who post their designs here seem to operate their 300B's at either 3-ish kOhm load (for higher power) or 5 kOhm load (for lower distortion). Most designers here seem to operate at around 350~400 V, 50~70 mA. A few minutes with the plate curves, a ruler, and a copy of RDH4 (specifically the section on SE output stages, pp 548~550) will convince the designer that a 300B needs to be operated near its max plate dissipation to get any power out at low distortion. My calculations claim 10.8 W @ 3.5 % 2nd harmonic (= 3.5 % THD roughly) for 3k, 400V, 100 mA. Furthermore, many commercial designs seem to operate at 3k, 400+ V, 90~100 mA.
At 4.2 kOhm, 370 V, 65 mA, I was getting 5.4 W into 8 Ohm @ 3 % THD (1 kHz) in my 300B amp.
Initially, increasing B+ resulting in glaring and harsh highs. This turned out to be my supply regulator that ran out of headroom. So I fixed the headroom issue and increased B+ to 410 V. This gives me 400 V between anode and cathode on the 300B. I cranked up the bias to 95~100 mA. With my Edcor CXSE25-8-5K (that actually measure 4.2 kOhm with 8 Ohm load - not the specified 5 kOhm -- measured using an HP 4194A Impedance Analyzer). With the new operating point, I get 12 W into 8 Ohm at a hair over 3 % THD (1 kHz). At 1 W it's 0.7 %.
It sounds really good too. And with nearly 25 W available total for a stereo amp, this think will rock the house off the foundation. It goes to at least eleven I think...
So why do people run 300B's at half plate dissipation? Tube life? Maybe if I was using $2k NOS WE 300B's I'd consider that. But for current production tubes?
~Tom
I'm puzzled by something. Why do most designers here run so wimpy operating points on their 300B's? Just askin'....
I've been designing this 300B amp for a while. Most people who post their designs here seem to operate their 300B's at either 3-ish kOhm load (for higher power) or 5 kOhm load (for lower distortion). Most designers here seem to operate at around 350~400 V, 50~70 mA. A few minutes with the plate curves, a ruler, and a copy of RDH4 (specifically the section on SE output stages, pp 548~550) will convince the designer that a 300B needs to be operated near its max plate dissipation to get any power out at low distortion. My calculations claim 10.8 W @ 3.5 % 2nd harmonic (= 3.5 % THD roughly) for 3k, 400V, 100 mA. Furthermore, many commercial designs seem to operate at 3k, 400+ V, 90~100 mA.
At 4.2 kOhm, 370 V, 65 mA, I was getting 5.4 W into 8 Ohm @ 3 % THD (1 kHz) in my 300B amp.
Initially, increasing B+ resulting in glaring and harsh highs. This turned out to be my supply regulator that ran out of headroom. So I fixed the headroom issue and increased B+ to 410 V. This gives me 400 V between anode and cathode on the 300B. I cranked up the bias to 95~100 mA. With my Edcor CXSE25-8-5K (that actually measure 4.2 kOhm with 8 Ohm load - not the specified 5 kOhm -- measured using an HP 4194A Impedance Analyzer). With the new operating point, I get 12 W into 8 Ohm at a hair over 3 % THD (1 kHz). At 1 W it's 0.7 %.
It sounds really good too. And with nearly 25 W available total for a stereo amp, this think will rock the house off the foundation. It goes to at least eleven I think...
So why do people run 300B's at half plate dissipation? Tube life? Maybe if I was using $2k NOS WE 300B's I'd consider that. But for current production tubes?
~Tom
There are many reasons for running tubes at different operating points, some of which you have touched on...
A lot of people like to run DHT's hot because they say they sound better, especially the Chinese / Russian versions. Some people prefer to run the tubes conservatively, to get more life out of them or they may like the sound at those lower operating points.
Some people run an 845 and only get 3~4 Watts out of it.
A lot of people like to run DHT's hot because they say they sound better, especially the Chinese / Russian versions. Some people prefer to run the tubes conservatively, to get more life out of them or they may like the sound at those lower operating points.
Some people run an 845 and only get 3~4 Watts out of it.
I use low dollar 300B's. I have not seen it with a Shuguang tube yet but a Sovtek 300B will go into runaway if you go over 400 volts despite the lettering on the box specifying 450 volts max. Not all of them will do it but 3 out of the 6 live ones that I have left will do it. The grid winding does not go far enough up in the tube so that there is an area at the top of the tube where there is a direct line of sight from the filament to the plate with no grid wires in between. Apply enough voltage and an uncontrolled current will flow between filament to plate even with -160 volts on the grid.
I have built a few 300B based Tubelab SE's including one I still have that runs Edcor CXSE 25-8-5K's. It runs about 380 volts of B+ and does sound best with the current cranked up to near 100mA. It makes about 8 watts. After zapping a Sovtek 300B into extinction and saving a second by yanking the cord when the glow started I figured out the avalanche effect and decided to keep all Sovteks below the 400 volt zone. I built a P-P 300B amp and also kept it right at 380 volts. Again it sounds best with the current cranked up. OPT is 6600 ohms.
All of this was done before Shuguang started making 300B's so I have not tried cranking them above 400 volts, but I have several of them. Most other 300B's are beyond my price range. I have heard one of my TSE's with WE 300B's and it sounded nice, but I can't say how much nicer than the Shuguangs that were in there.
Check the tubes that you run to see if you can cut the tube completely off at your chosen B+ voltage. ALL of the old Sovteks that I have will draw a few (5 to 10) mA at 400 volts even with -160 volts on the grid (limit of power supply). This current goes up as the plate voltage is increased and takes off rapidly once the "avalanche" threshold is reached.
And some of us get 40 WPC. I run the Chinese tubes at 100+ ma with 1050 volts on them. No glow, they ROCK! I have some old RCA's and GE's (got them cheap) but I tend to keep them below 70 mA since I am not going to buy more of them.
I have built a few 300B based Tubelab SE's including one I still have that runs Edcor CXSE 25-8-5K's. It runs about 380 volts of B+ and does sound best with the current cranked up to near 100mA. It makes about 8 watts. After zapping a Sovtek 300B into extinction and saving a second by yanking the cord when the glow started I figured out the avalanche effect and decided to keep all Sovteks below the 400 volt zone. I built a P-P 300B amp and also kept it right at 380 volts. Again it sounds best with the current cranked up. OPT is 6600 ohms.
All of this was done before Shuguang started making 300B's so I have not tried cranking them above 400 volts, but I have several of them. Most other 300B's are beyond my price range. I have heard one of my TSE's with WE 300B's and it sounded nice, but I can't say how much nicer than the Shuguangs that were in there.
Check the tubes that you run to see if you can cut the tube completely off at your chosen B+ voltage. ALL of the old Sovteks that I have will draw a few (5 to 10) mA at 400 volts even with -160 volts on the grid (limit of power supply). This current goes up as the plate voltage is increased and takes off rapidly once the "avalanche" threshold is reached.
And some of us get 40 WPC. I run the Chinese tubes at 100+ ma with 1050 volts on them. No glow, they ROCK! I have some old RCA's and GE's (got them cheap) but I tend to keep them below 70 mA since I am not going to buy more of them.
Last edited:
George,
I recall reading your comments and frustrations with the Sovtek 300B on your website. That's in part why I started conservative with the bias. Then I got to talking with Jack Elliano of Electra-Print and he strongly suggested crankin' up the voltage and current. And I do admit it's nice with the extra headroom.
I use JJ 300B tubes. I got them through AES at roughly $100/EA. The grid does go all the way through the anode sleeve at the top, but at the bottom there's a few mm of anode without grid. But the cathode doesn't go all the way to the bottom either. There's probably a bit of cathode without grid coverage. I have no issues cutting off the tube. At -150~-160 V Vgk, I get some 10 uA of anode current.
Thanks for your explanation of how to spot the onset of this "avalanche" phenomenon.
~Tom
I recall reading your comments and frustrations with the Sovtek 300B on your website. That's in part why I started conservative with the bias. Then I got to talking with Jack Elliano of Electra-Print and he strongly suggested crankin' up the voltage and current. And I do admit it's nice with the extra headroom.
I use JJ 300B tubes. I got them through AES at roughly $100/EA. The grid does go all the way through the anode sleeve at the top, but at the bottom there's a few mm of anode without grid. But the cathode doesn't go all the way to the bottom either. There's probably a bit of cathode without grid coverage. I have no issues cutting off the tube. At -150~-160 V Vgk, I get some 10 uA of anode current.
Thanks for your explanation of how to spot the onset of this "avalanche" phenomenon.
~Tom
I've been running JJ 300B for about 11yrs now, and run them at 400V + and typically 85mA - 90mA and have had no problems with them. I'm still running the quartet of original tubes, both sets having at least 5K hours on them.
WE 300B from experience will run reliably at these levels, but don't seem to last as long as the JJ based on the experience of all of my clients whose amps were originally delivered with WE 1999/2000 production. Can't speak to original production or any other time frame of production.
Svetlana 300B also ran fine at these voltage and current levels until filament sag killed most of them. I've avoided them since that time since I have no idea whether or not the issue was really satisfactorily resolved. FWIW The JJ sound much better IMHO than these do in SE applications, not such a big difference in PP.
Early Valve Arts and the even cheaper Shuguang tubes available to me at the end of the 1990s had the exact design issue George mentioned in his post on this issue. They also bluntly sound inferior to anything else I have tried.
The TJ stuff is of very good quality, and seems to run at reasonable voltages and currents, but the meshplate versions in particular should not be run too hard due to limited dissipation capability. I have very limited experience with these tubes, but hear little bad about them.
The current Shuguang Treasure series are supposed to be very good, but I have no experience with them at all, and have no real desire to find out.
WE 300B from experience will run reliably at these levels, but don't seem to last as long as the JJ based on the experience of all of my clients whose amps were originally delivered with WE 1999/2000 production. Can't speak to original production or any other time frame of production.
Svetlana 300B also ran fine at these voltage and current levels until filament sag killed most of them. I've avoided them since that time since I have no idea whether or not the issue was really satisfactorily resolved. FWIW The JJ sound much better IMHO than these do in SE applications, not such a big difference in PP.
Early Valve Arts and the even cheaper Shuguang tubes available to me at the end of the 1990s had the exact design issue George mentioned in his post on this issue. They also bluntly sound inferior to anything else I have tried.
The TJ stuff is of very good quality, and seems to run at reasonable voltages and currents, but the meshplate versions in particular should not be run too hard due to limited dissipation capability. I have very limited experience with these tubes, but hear little bad about them.
The current Shuguang Treasure series are supposed to be very good, but I have no experience with them at all, and have no real desire to find out.
Before spending money on the JJ tubes I read quite a few reviews of 300B tubes. The JJ's are viewed quite well by reviewers. I haven't tried any other brands of 300B's.
I wanted to run 400 V on the anode (cathode grounded), but my supply runs out of headroom at around 390 V on low AC supply (114 V). I set it at 385 V which leaves about 375 V across the 300B. It measures well and sounds sweet.
With just one channel running, I can get the supply to run at 400 V. I'm not noticing any difference in the sound quality between it and 385 V. It measures slightly better at the higher voltage - 0.60 % THD @ 400 V, 100 mA, 1 kHz, 1 W vs 0.66 % THD @ 385 V, 100 mA, 1 kHz, 1 W. The max power drops by maybe 0.25~0.5 W at the lower voltage. I kinda went meh.... Changing to a higher voltage would result in a significant redesign loop as I'm using a regulated supply and need to ensure that the front end of the supply can survive running without load at high AC mains voltage.
~Tom
I wanted to run 400 V on the anode (cathode grounded), but my supply runs out of headroom at around 390 V on low AC supply (114 V). I set it at 385 V which leaves about 375 V across the 300B. It measures well and sounds sweet.
With just one channel running, I can get the supply to run at 400 V. I'm not noticing any difference in the sound quality between it and 385 V. It measures slightly better at the higher voltage - 0.60 % THD @ 400 V, 100 mA, 1 kHz, 1 W vs 0.66 % THD @ 385 V, 100 mA, 1 kHz, 1 W. The max power drops by maybe 0.25~0.5 W at the lower voltage. I kinda went meh.... Changing to a higher voltage would result in a significant redesign loop as I'm using a regulated supply and need to ensure that the front end of the supply can survive running without load at high AC mains voltage.
~Tom
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.