The grid is at -113v.
I was thought that they could be adjusted independently. I can lower the B+ as you suggest (the Maida regulator is adjustable) so I’ll give that a try. I took the OPTs off so it will take a couple of days though.
Thanks for your input!
I was thought that they could be adjusted independently. I can lower the B+ as you suggest (the Maida regulator is adjustable) so I’ll give that a try. I took the OPTs off so it will take a couple of days though.
I was thinking to keep the B+ on the 660v tap and feed the B- from the 520v.
Thanks for your input!
Regarding lowering the B+ with the Maida, I'm not sure it's going to work the way you think it will but to be honest, I don't know. As I understand it and using the amp without the Maida, if you start with very low current through the output tube, the B+ will be low and the B- very negative. As you increase the current, B+ increases and B- becomes less negative. My thought was reducing B+ with a resistor might allow B+ to start out a little lower, then as you increased the 300b current you would have a little more room to make B- less negative before you ran into too much B+ and current. If you put a Maida in the system, it's an active device acting like a voltage regulator. I don't know what an active device might do but my gut tells me that it will be fighting against your bias adjustments as it tries to regulate B+. I probably am misunderstanding this.
I don't know enough to say so, but I don't think you can use asymmetric taps (66V on B+ and 520v on B-) on the transformer. I'm guessing that it would provide a DC offset in the transformer which would cause it to overheat.
Since you have R4 in the system, it would make sense to go back to the beginning and do the amp setup as outlined in the Assembly manual. That should get you to a workable baseline where you can try the amp as designed. Then start playing with the Maida and so on. Given that your grid voltage is very close to New Sensor's negative voltage limit and that it's almost double the normal range, I'm not surprised that you are not real happy about how it sounds.
Best of luck in your next steps.
I don't know enough to say so, but I don't think you can use asymmetric taps (66V on B+ and 520v on B-) on the transformer. I'm guessing that it would provide a DC offset in the transformer which would cause it to overheat.
Since you have R4 in the system, it would make sense to go back to the beginning and do the amp setup as outlined in the Assembly manual. That should get you to a workable baseline where you can try the amp as designed. Then start playing with the Maida and so on. Given that your grid voltage is very close to New Sensor's negative voltage limit and that it's almost double the normal range, I'm not surprised that you are not real happy about how it sounds.
Best of luck in your next steps.
Hi Lehman - just to summarise: the board is set up as in the manual. The Maida regulator just goes in place of R4 and C5 is reduced to 10uf. I have run the board with R4 (exactly as in the manual) and with the regulator and get the same results. The assembly manual suggests running 300b at around 390v and 70mA which is what I am doing. I couldn’t find any info on what to expect for the B- when running at 390v. The setup description in the manual says that at at 285v B+ to expect around -185v B-. I think that the lower B- that I have now is what one would expect when using 660v taps. Do you have more info on New Sensors negative voltage? I haven’t heard of that before.
There is a lot of bass distortion so I am trying to troubleshoot why that is happening. It may be my speakers as I built them based around another amp and the woofers are likely to be quite high efficiency. I will test the system with REW to see if the bass distortion may be because of the woofers.
Thanks!
Sorry. I didn't understand how the Maida worked in this system.
New Sensor is the maker of Electro-Harmonix. I didn't think to just use E-H. Here is a link to their 300b datasheet.
https://shop.ehx.com/catalog/tubes-specs/300B gold.pdf
This data sheet has limits for the grid voltage. Notice in the graph at the bottom that they only take it to a plate voltage of 350V and a grid voltage of -80V. I agree with you that a B- of -250V is probably OK and a product of a higher B+. I think that grid voltage relates more to the curves on the datasheets and maybe more useful, at least to me.
The other datasheet that I found interesting is from Emission Labs.
http://www.emissionlabs.com/datasheets/EML300B.html
They show a table of Anode to Heater (essentially B+), grid to heater (grid to ground), and anode or plate current. For a 400V B+, grid voltage should be in the negative mid-80s. That seems to point to something wrong in your system.
Also, the curve at the bottom suggests that for -113V grid and 60 mA plate current you should have well over 500V B+. For your actual B+ and current, you should be between -75 and -90 for grid voltage. Could it be the tubes? I don't know. If you have a Variac you could slowly increase your source voltage using your good tubes while making sure to reset the bias often and staying within normal B+ and grid voltage.
New Sensor is the maker of Electro-Harmonix. I didn't think to just use E-H. Here is a link to their 300b datasheet.
https://shop.ehx.com/catalog/tubes-specs/300B gold.pdf
This data sheet has limits for the grid voltage. Notice in the graph at the bottom that they only take it to a plate voltage of 350V and a grid voltage of -80V. I agree with you that a B- of -250V is probably OK and a product of a higher B+. I think that grid voltage relates more to the curves on the datasheets and maybe more useful, at least to me.
The other datasheet that I found interesting is from Emission Labs.
http://www.emissionlabs.com/datasheets/EML300B.html
They show a table of Anode to Heater (essentially B+), grid to heater (grid to ground), and anode or plate current. For a 400V B+, grid voltage should be in the negative mid-80s. That seems to point to something wrong in your system.
Also, the curve at the bottom suggests that for -113V grid and 60 mA plate current you should have well over 500V B+. For your actual B+ and current, you should be between -75 and -90 for grid voltage. Could it be the tubes? I don't know. If you have a Variac you could slowly increase your source voltage using your good tubes while making sure to reset the bias often and staying within normal B+ and grid voltage.
Thanks again. I could try lowering b+, but I think it will be hard to get b- to above -100v. Could I also change r6 to bring up b-? Isn’t b- value set by the ratio between r11-13 and r6? I am thinking of trying to raise r6 from 10k to 50k. But I remember George said that B- needs to be running before B+, and raising r6 would slow b-. I really appreciate your help!
The TSE died a long time ago when the Sharp filament regulator went extinct due to Foxconn's takeover of Sharp Electronics. There were plenty of regulators in stock at the usual distributors and it seemed that TSE builders were the only people buying them. One day ALL of the chips vanished, and one Ebay seller was offering them for like $25 each. I had purchased a couple of tubes full back when I designed the TSE, so I sent a free chip with each board while I cooked up the TSE-II. Unfortunately I ran out of chips before the TSE-II was done, but lapse was brief. The heart of the TSE lives on inside the TSE-II.
The TSE-II has the exact same audio path as the TSE, but the filament regulator was replaced with a Microchip part, and the board was made larger with some added holes to improve the air flow around the mosfets. The original TSE did run rather hot since I originally intended it for 45 tubes. Several requests were discussed at the beginning of this thread, and most were accommodated.
One of the most often requested change was to ditch the 5842 in favor of something more available. 5842's were cheap and common when I designed the TSE in 2003 / 2004. It was the best tube for the job 20+ years ago but they are not so easy to find in Europe or Australia today. Will there be a TSE-III, and an UNSET-II?
While attempting to invent the Shredmaster 5000 guitar amp, I have accidentally invented a new driver circuit that produces enough drive voltage for 300B's at THD levels in the under 0.1% range! The best part is that there are about 40 different 7 pin tubes that work in it (some for $1 each), and 15 different 9 pin tubes that work. I am leaning toward the 7 pin tubes, but testing everything.
Did the amp work correctly before adding the Maida regulator? I always build things as close to the original design as possible, get it working, then mod it. This applies to everything from car engines to my own boards. If you build and mod at the same time, and it doesn't work, there are two possible unknowns that work against you.
Raising R6 to 15k or even 20K will not make the B- slower than a 5AR4, so you should be OK, but I have never needed to mess with it.
First off, ignore the actual B- voltage. Excess voltage may make R14 and R25 get a bit warmer, but it won't affect the operation of the amp. Excess voltage actually allows for more headroom in the drive circuit. The positive and negative voltages serve to power the mosfets. Pots R12 and R23 set the gate voltage on the mosfets which drive the grids of the output tubes. As long as the tube current can be set to about 70 mA with adjustment room to hit 100, there should be no issue with the amp's operation. Verify that there is about 150 volts across D6. If not, there will be clipping on strong peaks.
I looked back at my pictures and counted 6 Tubelab amps I have built, including a TSE-II that I recently sold. Before retiring, several students bought different Tubelab boards and built them. Every time they played music through their Tubelab amp for the first time, I would yell: it's alive! It's Alive! IT'S ALIVE!!! As loud as I could. The students always were amazed at what they had learned and done. There are several Tubelab amps still out there in various locations of the world, thanks to Tubelabs. And, I still have about 10 Sharp filament regulators, even though they went extinct due to Foxconn's takeover of Sharp Electronics, but I'm not the person who ebay selling them of $25/each.
One student built this push pull amp based on a Tubelab board:
One student built this push pull amp based on a Tubelab board:
I sure hope.
I have squirreled away 6 pairs of 5842s, but that’s not much insurance.
Hi George, thanks for the reassurance. I took out the Maida, and I have r4 in with 405v upstream and 397v for b+. The mosfet voltage across d6 is 160v and b- is -245v. There is 70mA on each 300b and the 5842s are at 170v. The grid is -97v. I think it might be clipping as the effect gets much worse at higher volumes. I will try now to post a recording. Henry
Last edited:
Too big for diyaudio so I posted on YouTube. Here is the my reference amp:
Here is the Tse:
Thanks!
It sounds like boz scaggs has a frog in his throat and the bass is wobbly. Sorry I am probably wasting everyone’s time, I am guessing that I have done something basic wrong. 😑
Hi elarte, it is in both channels. I also tried a spare set of valves, and OPTs.
That sounds like a low level hum, or hum intermodulating with the music signal. Two possibilities come to mind. My experiments, and those of other people revealed that the output tubes did not like being remote mounted on the ends of long wires. Having a mosfet directly connected to the grid of a tube creates a sub-circuit with a lot of Gm, which makes oscillation at RF frequency a possibility. See if the amp has any effect on a portable over-the-air radio or TV set. If so the oscillation muse be cured.
The low level growl that is present in the music seems to be at a constant frequency, possibly related to the power line frequency. This can occur if the filament regulator is in, or near dropout. Try this simple test. remove line power from the amp while music is playing. The amp should continue to play for a second or two. Does the music sound clear during that second? Note, Do not re-apply power immediately, wait a minute or two. Rectifier tubes don't like a brief interruption in power.
Measure the filament voltage on the 300B's, should be close to 5 volts. Measure the voltage going into the regulator chip (end of R3 closest to rectifier tube). It should be over 6 volts. If you have a scope, look at the filament voltage on the 300B's. It should be ripple free DC.
The low level growl that is present in the music seems to be at a constant frequency, possibly related to the power line frequency. This can occur if the filament regulator is in, or near dropout. Try this simple test. remove line power from the amp while music is playing. The amp should continue to play for a second or two. Does the music sound clear during that second? Note, Do not re-apply power immediately, wait a minute or two. Rectifier tubes don't like a brief interruption in power.
Measure the filament voltage on the 300B's, should be close to 5 volts. Measure the voltage going into the regulator chip (end of R3 closest to rectifier tube). It should be over 6 volts. If you have a scope, look at the filament voltage on the 300B's. It should be ripple free DC.