Jhofland has designed a nice diamond buffer that I will be offering as a group buy via a preorder through my shop. It’s an all BJT design fully preassembled in a very compact SMT format about the size of a postage stamp. I have been testing it and listening to it for about a week now. It sounds superb and even has enough grunt to drive 32ohm planar HiFiman HE-400i headphones. It can be used as a front end on many projects, limited only by your imagination (M2X, F6, etc). The buffer has been tested with +/-20Vdc rails and should work fine with +/-24v rails or even a bit higher. I’ll know more with added testing. The technical discussion for this buffer is here:
https://www.diyaudio.com/community/threads/jhoflands-diamond-buffer.401231/#post-7400252
I have found that the BTSB or BTSB hybrid tube (HyperSET) buffer/preamp boards make for an excellent voltage gain stage for this buffer.
This thread is to discuss the interest for a GB for a fully assembled and ready to run (RTR) Diamond buffer module.
Here is the schematic:
Here is render of the PCB assembly:
Here are samples under test as a high performance headphone amp output stage. Here, the diamond buffer is being driven by the HyperSET tube buffer set to 14dB gain and the +/-20v PSU is provided by the Yarra Mk I linear PSU with active bridges, a CLC, and TI TPS7A4XXX low noise regulators (one of the best linear preamp/headphone amp PSU’s that I have):
Edit Nov 24, 2023 - Black Friday: The PCBA’s arrived and I have tested them and they work well but I discovered that I mistakenly had the board house install a red LED in place of a green LED (this was one of the parts they substituted that I did not catch). The red LED causes the bias current through the diamond input stage to be about 40% below the 1mA nominal value. It works fine for most applications but if you want to drive high voltages into lower impedance loads, you could swap the LED out with a hot air pencil). But as I said the board works fine but is not perfect and still very useable and nice sounding, so in the spirit of the holiday season, and following what we did a few years ago with the Xmas Amp, I will offer these diamond buffers fully built and RTR for FREE (while supplies last). You will need to pay for shipping (auto-calculated from my Etsy shop). I will put the link to order shortly.
Operationally, you just need to self install the 2-pin and 4-pin JST or Molex KK connectors of your choice (or solder flying leads if you wish). Provide a suitable dual rail supply, connect your audio input and you have a nice output that can drive just about any preamp type load or headphones.
Boards will be 1oz copper ENIG finish with green solder mask. Sold as fully pre-assembled pairs with high quality genuine parts. You only need to install the through hole input and output connectors of your choice.
A big thanks to Jhofland for designing this superb circuit.
BOM here:
https://www.diyaudio.com/community/threads/jhoflands-diamond-buffer-gb.401244/post-7517251
https://www.diyaudio.com/community/threads/jhoflands-diamond-buffer.401231/#post-7400252
I have found that the BTSB or BTSB hybrid tube (HyperSET) buffer/preamp boards make for an excellent voltage gain stage for this buffer.
This thread is to discuss the interest for a GB for a fully assembled and ready to run (RTR) Diamond buffer module.
Here is the schematic:
Here is render of the PCB assembly:
Here are samples under test as a high performance headphone amp output stage. Here, the diamond buffer is being driven by the HyperSET tube buffer set to 14dB gain and the +/-20v PSU is provided by the Yarra Mk I linear PSU with active bridges, a CLC, and TI TPS7A4XXX low noise regulators (one of the best linear preamp/headphone amp PSU’s that I have):
Edit Nov 24, 2023 - Black Friday: The PCBA’s arrived and I have tested them and they work well but I discovered that I mistakenly had the board house install a red LED in place of a green LED (this was one of the parts they substituted that I did not catch). The red LED causes the bias current through the diamond input stage to be about 40% below the 1mA nominal value. It works fine for most applications but if you want to drive high voltages into lower impedance loads, you could swap the LED out with a hot air pencil). But as I said the board works fine but is not perfect and still very useable and nice sounding, so in the spirit of the holiday season, and following what we did a few years ago with the Xmas Amp, I will offer these diamond buffers fully built and RTR for FREE (while supplies last). You will need to pay for shipping (auto-calculated from my Etsy shop). I will put the link to order shortly.
Operationally, you just need to self install the 2-pin and 4-pin JST or Molex KK connectors of your choice (or solder flying leads if you wish). Provide a suitable dual rail supply, connect your audio input and you have a nice output that can drive just about any preamp type load or headphones.
Boards will be 1oz copper ENIG finish with green solder mask. Sold as fully pre-assembled pairs with high quality genuine parts. You only need to install the through hole input and output connectors of your choice.
A big thanks to Jhofland for designing this superb circuit.
BOM here:
https://www.diyaudio.com/community/threads/jhoflands-diamond-buffer-gb.401244/post-7517251
Last edited:
re setting the output. You don't. The nature of this circuit is that the output DC level follows the input. Of the 10 prototypes I tested at +/-15V supply and input shorted the worst offset was 6mV. Average was about 4mV.
re: signal level. It depends on your supply rails. Peak output is about Vsupply -3V. Nominal output impedance is between 6 and 7 ohms. I did not test this circuit to destruction I expect the output power will depend on the ability for the output devices to withstand the thermal considerations.
re: signal level. It depends on your supply rails. Peak output is about Vsupply -3V. Nominal output impedance is between 6 and 7 ohms. I did not test this circuit to destruction I expect the output power will depend on the ability for the output devices to withstand the thermal considerations.
I measured 4.5mV DC offset on one channel and -1.5mV on the other one. No trimpot needed anywhere as Jhofland says, the output DC level follows the input.
X,
Can you show the THD graph with a typical headphone load, i.e. 32 ohms and 300 ohms?
Also:
Can you show the THD graph of just the buffer (sans HyperSET) into a typical 10K ohm load? This would be more apropos if you are marketing towards users replacing 0dB gain input buffers of various solid state amp circuits. You can also show how it fairs at various input voltages (1V, 2V, 4V, 8V, etc...)
Thanks,
Anand.
Can you show the THD graph with a typical headphone load, i.e. 32 ohms and 300 ohms?
Also:
Can you show the THD graph of just the buffer (sans HyperSET) into a typical 10K ohm load? This would be more apropos if you are marketing towards users replacing 0dB gain input buffers of various solid state amp circuits. You can also show how it fairs at various input voltages (1V, 2V, 4V, 8V, etc...)
Thanks,
Anand.
Hi Anand,
Working on the 32ohms - as designed, the bias current is not high enough to support clean driving of 32ohm loads so I am working on LTSpice to determine new setpoint and will test a revised low impedance version. I think the current board is fine for a buffer/preamp but not ideal for a 32ohm headphone amp.
The THD shown with HyperSET is actually less than with plain HyperSET driving the 2k2 load.
I’ll change the sunny load to 10k and drive directly Victor’s part per billion THD oscillator. Although that does not have very good fine control of amplitude. I think the limiting step will be the Focusrite’s intrinsic THD.
Working on the 32ohms - as designed, the bias current is not high enough to support clean driving of 32ohm loads so I am working on LTSpice to determine new setpoint and will test a revised low impedance version. I think the current board is fine for a buffer/preamp but not ideal for a 32ohm headphone amp.
The THD shown with HyperSET is actually less than with plain HyperSET driving the 2k2 load.
I’ll change the sunny load to 10k and drive directly Victor’s part per billion THD oscillator. Although that does not have very good fine control of amplitude. I think the limiting step will be the Focusrite’s intrinsic THD.
Hi Anand,
Working on the 32ohms - as designed, the bias current is not high enough to support clean driving of 32ohm loads so I am working on LTSpice to determine new setpoint and will test a revised low impedance version. I think the current board is fine for a buffer/preamp but not ideal for a 32ohm headphone amp.
The THD shown with HyperSET is actually less than with plain HyperSET driving the 2k2 load.
I’ll change the dummy load to 10k and drive it directly with Victor’s part per billion THD oscillator. Although that does not have very good fine control of amplitude. I think the limiting step will be the Focusrite’s intrinsic THD.
Working on the 32ohms - as designed, the bias current is not high enough to support clean driving of 32ohm loads so I am working on LTSpice to determine new setpoint and will test a revised low impedance version. I think the current board is fine for a buffer/preamp but not ideal for a 32ohm headphone amp.
The THD shown with HyperSET is actually less than with plain HyperSET driving the 2k2 load.
I’ll change the dummy load to 10k and drive it directly with Victor’s part per billion THD oscillator. Although that does not have very good fine control of amplitude. I think the limiting step will be the Focusrite’s intrinsic THD.
The boards arrived from the PCB assembly house. They look great and sound great. However there is a red LED in place of a green one. This was a substitution the assembly house made that I did not catch. It means the bias current in the diamond input stage is about 40% below the nominal 1mA. But it works well for most applications unless you need it to drive high voltage swings into low impedance loads. It’s easy to swap out the red LED with a hot air pencil if you want. The boards are still very functional and a nice buffer to have on hand. But due to the mistake, they are not perfect so I will give them away for Free as the Xmas Diamond Buffer while supplies last (1 pair per member, please). You will need to pay for shipping (auto calculated per my Etsy shop). It’s appropriate that the LED glows red in color for the holidays! 
Last edited: