Hi,
Long time lurker, but this is my first post here. I'm thinking of building a pmillett push-pull 300b amp using his universal pcbs Push-pull KT88 class A amp with "universal" driver PCB.
It doesn't seem like anyone has done this before (or at least I can't find anyone who has done this on the web). This is my first tube amp build so I want to make sure I'm not doing anything really stupid. I have a few questions:
1) I plan to run the 300b filaments off 5V DC using Pete Millett's filament supply boards. Is there any reason why I shouldn't?
2)Do I need to lower the B+ voltage? From the 300b specs, the 300b can take 400V which is over the data given on Pete's website. More specifically, the power transformer I have also has 320V taps (in addition to 390V) taps. Would it better to use the 320V taps?
3)I connect the cathode output off the board to pin 1 on the 300b right?
4) If my preamp's output is already balanced (from a centre-tapped transformer I think) can I omit the input transformers?
Thanks in advance for all your help!
Long time lurker, but this is my first post here. I'm thinking of building a pmillett push-pull 300b amp using his universal pcbs Push-pull KT88 class A amp with "universal" driver PCB.
It doesn't seem like anyone has done this before (or at least I can't find anyone who has done this on the web). This is my first tube amp build so I want to make sure I'm not doing anything really stupid. I have a few questions:
1) I plan to run the 300b filaments off 5V DC using Pete Millett's filament supply boards. Is there any reason why I shouldn't?
2)Do I need to lower the B+ voltage? From the 300b specs, the 300b can take 400V which is over the data given on Pete's website. More specifically, the power transformer I have also has 320V taps (in addition to 390V) taps. Would it better to use the 320V taps?
3)I connect the cathode output off the board to pin 1 on the 300b right?
4) If my preamp's output is already balanced (from a centre-tapped transformer I think) can I omit the input transformers?
Thanks in advance for all your help!
I have no experience with Pete's circuits, but from DIY Audio chatter, he seems to have his stuff together so I don't see any reason why they wouldn't work.
One 300B PP amplifier I've been drooling over is Electra-Print's Push-Pull A2 Design. I love the simplicity. I'd probably use something more modern than the TDA2030, though. Maybe a couple of LME49600 in parallel or something.
One thing with 300B's is that you need to be careful when you operate them close to the max plate voltage. Some of them won't shut off completely under those operating conditions even when the grid is driven strongly negative. This leads to an avalanche-like run-away which destroys the tube. So if you operate under those conditions, make sure by turning the bias pot that you can turn the tube current completely off. I have had no issues with the JJ 300B's I'm using in my 300B SE amp at 400~410 V, 90~100 mA.
Heater supply: Definitely go with DC here. There are many options. I looked at the various options and concluded that under worst case conditions for a linear voltage regulator I'd burn more power in the regulator heat sink than was delivered to the speakers. I found this rather ridiculous. So I started a Universal Filament Regulator project using a switchmode regulator. Specifically for one of my 300B amps, I used an LMZ12002. That worked quite well.
~Tom
One 300B PP amplifier I've been drooling over is Electra-Print's Push-Pull A2 Design. I love the simplicity. I'd probably use something more modern than the TDA2030, though. Maybe a couple of LME49600 in parallel or something.
One thing with 300B's is that you need to be careful when you operate them close to the max plate voltage. Some of them won't shut off completely under those operating conditions even when the grid is driven strongly negative. This leads to an avalanche-like run-away which destroys the tube. So if you operate under those conditions, make sure by turning the bias pot that you can turn the tube current completely off. I have had no issues with the JJ 300B's I'm using in my 300B SE amp at 400~410 V, 90~100 mA.
Heater supply: Definitely go with DC here. There are many options. I looked at the various options and concluded that under worst case conditions for a linear voltage regulator I'd burn more power in the regulator heat sink than was delivered to the speakers. I found this rather ridiculous. So I started a Universal Filament Regulator project using a switchmode regulator. Specifically for one of my 300B amps, I used an LMZ12002. That worked quite well.
~Tom
There is more to what you are proposing than may be evident to you.
The first thing you need to determine is the operating conditions of the 300Bs. Without this information you don't know how much voltage drive you need. I run my 300Bs very conservatively based upon my desire to extend the life of the expensive tubes, but just as important, to run the tubes in the sweet spot for lowest distortion. The chart WE provided shows the lowest 3rd harmonic distortion at Vp = 350V, Vg = -76V, @ 50mA plate current for a power output of about 12W for the PP pair. So you see this operating point is conservative and has low distortion.
Personally I run my 300Bs at Vp = 350, Vg = -74 @ 60mA into a 3k ohm load.
For this operating point you will need a voltage drive of at least 74V peak and for 3db of headroom you will need about 105V drive. I seriously doubt that P. Millet's board can handle this, but you will need to ask him about it. Don't forget that this board is intended for use with a KT88 which needs about 35V of drive or 50V with 3dB of headroom.
Don't forget that the cathode resistor is subtracted from the B+ voltage so that for a 350V plate voltage you will need a B+ of 425V.
I hope you see that this is not a mix and match bolt together project. You will have to do some engineering.
The first thing you need to determine is the operating conditions of the 300Bs. Without this information you don't know how much voltage drive you need. I run my 300Bs very conservatively based upon my desire to extend the life of the expensive tubes, but just as important, to run the tubes in the sweet spot for lowest distortion. The chart WE provided shows the lowest 3rd harmonic distortion at Vp = 350V, Vg = -76V, @ 50mA plate current for a power output of about 12W for the PP pair. So you see this operating point is conservative and has low distortion.
Personally I run my 300Bs at Vp = 350, Vg = -74 @ 60mA into a 3k ohm load.
For this operating point you will need a voltage drive of at least 74V peak and for 3db of headroom you will need about 105V drive. I seriously doubt that P. Millet's board can handle this, but you will need to ask him about it. Don't forget that this board is intended for use with a KT88 which needs about 35V of drive or 50V with 3dB of headroom.
Don't forget that the cathode resistor is subtracted from the B+ voltage so that for a 350V plate voltage you will need a B+ of 425V.
I hope you see that this is not a mix and match bolt together project. You will have to do some engineering.
I think it will work to drive 300B grids:
Some specs for the driver section:
@ B+ = 375V:
Vout max at clipping is 253V P-P (+/-126V)
Vout for THD=5% into 100k load is 70V RMS, or 197V P-P
THD @ 10V RMS out is ~0.24%
THD @ 20V RMS out is ~0.55%
Vout for input of 100mV (no negative feedback) is 10.77V RMS, or a gain of about 100
...but I never tried it myself.
I'm sure some optimization could be done as well.
Pete
Some specs for the driver section:
@ B+ = 375V:
Vout max at clipping is 253V P-P (+/-126V)
Vout for THD=5% into 100k load is 70V RMS, or 197V P-P
THD @ 10V RMS out is ~0.24%
THD @ 20V RMS out is ~0.55%
Vout for input of 100mV (no negative feedback) is 10.77V RMS, or a gain of about 100
...but I never tried it myself.
I'm sure some optimization could be done as well.
Pete
I've found that the JJ, TJ (mesh) and WE 300B will all operate safely at 400V and from 70mA - 90mA based on experience with my commercial SE and PP 300B amp designs. (I would not run the mesh plates beyond 70mA) I can also confirm Tom's comments on the avalanche run away issues that some cheaper 300Bs exhibit.
Short answer no, the input transformer is used as the phase splitter in this design.
However if your pre-amp indeed has a center tap on it's output transformer you can use, then you might be able to do some double duty with it...
Pete's Push-Pull board should work fine for this. But it may be better to try it with KT88s first - that way you can follow Pete's design, and try to hear the sound that the board was designed to make.
300B output DHTs can sound wonderful, but you must go to a lot of trouble to get any advantage over the original KT88 design.
Start with output transformer (the most important choice in the amp). You need an OT of highest quality - this is far more important than tube choice. Low cost OT with 300B makes no sense. Similarly, B+ power supply must be highest quality, preferably choke-input. For 420 to 440V dc B+ (required for cathode-bias amps: 350V anode voltage, + 70-90V of cathode bias voltage) then the power transformer secondary must be about 460V rms for choke-input design. Making a B+ using raw dc with no choke, and using electrolytic capacitors is not good enough, and again 300B will give no real improvement over KT88.
Filament supply. KT88 can operate with ac - very easy. In PP, 300B can also work with ac, and the humm may not be too bad. But the filament is the cathode in a 300B, so ac heating will add 100/120Hz sidebands to the music signal - and the ac-transformer will conduct mains-noise into the cathode.
Many dc solutions are even worse sounding than ac though. Again, this is because the filament is the cathode of your tube, and directly connected to the music. Rectified dc will conduct rectifier recovery pulses right into your music signal. Also, note that the filament has two terminals, which means 2 cathode terminals. The filament heating voltage makes one terminal 5V higher than the other - so it is 5V closer to the anode. This is very important for the filament supply design, because it means that the music signal can flow into the filament regulator. If you have a dc solution that presents an electrolytic capacitor across the filament, then the music signal flows into an electrolytic capacitor, in a similar way to using an electrolytic cathode bypass capacitor. This can be a limitation on the performance of an amplifier - especially when you eliminate electrolytics elsewhere in the amp.
Solutions for filament heating that address the problems of shielding the filament from external noise, and preventing audio signal circulating in the filament supply include using large chokes, as Thomas Mayer does, and solid-state solutions from Tentlabs (ready assembled) or as kits from me! No need for me to talk about the difference these make to the sound - you can see what DIYers think, after they have compared different solutions:
http://www.diyaudio.com/forums/tubes-valves/38248-new-dht-heater-5.html#post2336382
.
^ Thanks for the detailed post.
I will be using a Tamura F-2021 5K Push-Pull transformer which a should be pretty good quality and a Tamura choke. I also already have a quad of TJ Full Music 300b/n so with regards to tubes I am ready to go.
On the other hand, I already have a quad of NOS Tung-Sol 6550 grey solid plate. Would I be better off sticking with these rather than 300b? One of the main reasons I want to go 300b is my pre-amp is a 300b SET.
I will be using a Tamura F-2021 5K Push-Pull transformer which a should be pretty good quality and a Tamura choke. I also already have a quad of TJ Full Music 300b/n so with regards to tubes I am ready to go.
On the other hand, I already have a quad of NOS Tung-Sol 6550 grey solid plate. Would I be better off sticking with these rather than 300b? One of the main reasons I want to go 300b is my pre-amp is a 300b SET.
PP 300B amplifiers have a different character to SE 300B. If you want a similar character of sound, maybe 300B-SE will work better. If you listening room has a floor area of 3m x 3m up to maybe 5m x 5m, and your speakers are efficient to 92dB/W or more, 300B-SE is wonderful, especially at low volumes.
As for the PP build choices, you make it easy!
If you have high quality 6550 and 5K OT, then you can build the Pete Millet PP board according to the instructions, and you'll have a known-good sounding design.
It's really easy to make a bad-sounding amp with directly-heated triodes.
But, when you know how good the 6550 version can be, you can introduce the 300B components, beginning with ac filaments. With PP, the 300Bs won't hum too bad (unlike with 300B-SE), so you can see if 300B can sound as good as 6550.
If you can make it sound good with 300B, you can upgrade the filament heating to a high-quality dc solution, and really hear what directly-heated triodes can do. This should be the standard test for all DHT filament supplies: if it's not HUGELY better, in every way, compared to ac-heating, then it's not good enough.
As for the PP build choices, you make it easy!
If you have high quality 6550 and 5K OT, then you can build the Pete Millet PP board according to the instructions, and you'll have a known-good sounding design.
It's really easy to make a bad-sounding amp with directly-heated triodes.
But, when you know how good the 6550 version can be, you can introduce the 300B components, beginning with ac filaments. With PP, the 300Bs won't hum too bad (unlike with 300B-SE), so you can see if 300B can sound as good as 6550.
If you can make it sound good with 300B, you can upgrade the filament heating to a high-quality dc solution, and really hear what directly-heated triodes can do. This should be the standard test for all DHT filament supplies: if it's not HUGELY better, in every way, compared to ac-heating, then it's not good enough.
Thanks, Rod, for your wise words. I'll keep them in mind however, I have a gut feeling that a 300b PP will "mesh" better with a SE300B preamp than a KT88/6550 PP.
^ Isn't the output of my preamp is already phase-split by its output transformers since it has XLR outs (only).
^ Isn't the output of my preamp is already phase-split by its output transformers since it has XLR outs (only).
Okay so I ended up building a 300b push pull with Pete Millett's DC filament boards. Here are some pics:
I didn't use an input transformer to save some money.
It sounds great but I'm lacking just a little gain. Will adjusting the cathode bias resistor increase the gain -- I'm using 400V to supply the boards and 100ohm cathode bias resistors. Alternatively my transformer has a 320V tap -- will lowering to 320V increase the gain if I keep the same cathode bias resistor?
Otherwise, what do I need to do to increase the gain?
I didn't use an input transformer to save some money.
It sounds great but I'm lacking just a little gain. Will adjusting the cathode bias resistor increase the gain -- I'm using 400V to supply the boards and 100ohm cathode bias resistors. Alternatively my transformer has a 320V tap -- will lowering to 320V increase the gain if I keep the same cathode bias resistor?
Otherwise, what do I need to do to increase the gain?
Last edited:
Thanks!
I'm lucky enough to be able to get chassis like this locally so I'm afraid I can't help you there!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.