Then bias output stage as close as possible to the zero tempco region for the device you are using, then aim to achieve zero tempco for the front end.
Maybe use laterals on the output, biased at 100mA.
I am planning to use some good old 2SK216 laterals as outputs, which I found in my parts bin. Since I am designing this for 16 ohm headphones I want a bit more bias, maybe 250 mA or so. However, as far as I understand it, this bias is higher than the zero tempco point (datasheet attached). I'll have to figure out another way to keep warmup drift under control. View attachment rej03g0903_2sk213ds_a.pdf
Don't worry.
250mA will be fine, even far away from the zero tempco they hardly drift at all.
Just leave the front end as is no need for leds, or maybe use lm329 instead if you really hated zeners.as it will kind of cancel each other out.
You could always use double die laterals etc, many ways to skin this cat.
250mA will be fine, even far away from the zero tempco they hardly drift at all.
Just leave the front end as is no need for leds, or maybe use lm329 instead if you really hated zeners.as it will kind of cancel each other out.
You could always use double die laterals etc, many ways to skin this cat.
Last edited:
Hmmm, I am trying to get ready to start building my version of the Aleph J for headphones. I think I sort of understand most of how the circuit works. However, I am not sure about the source resistor for the gain device (R35 in the schematic shown in the first post of this thread). I see different values for this in the different versions of the Aleph circuit. I guess this has to do with (1) the bias current and (2) the output device -- but I don't know how to wrap my head around this. How do I choose a suitable value for this resistor?
P.S.: I am aiming for a bias current of 300 mA with 2SK214 output devices.
P.S.: I am aiming for a bias current of 300 mA with 2SK214 output devices.
Ok, after a little bit of tweaking and some help from ZM I got the prototypes of my headphone version of the Aleph-J working pretty well. It goes up to 100 kHz (-3dB), clipping starts at 300 mW into a 16 Ohm load (ears would be gone long before that). Initial listening is good!
I am waiting to get the last parts to finalise the PSU. Once I have everything, I will think about a proper chassis for this.
I am waiting to get the last parts to finalise the PSU. Once I have everything, I will think about a proper chassis for this.
Attachments
It has been quite a few since I last worked on my Aleph H. Although the last version with lower bias worked fine, I remembered (or at least thougt I did) that it sounded better with the (ridiculous) 400-500mA bias I had previously tried, when I had trouble taming the hum. I would need an enclosure with real heatsinks if I wanted to resurrect that version!
I bought a small enclosure from aliexpress (see photo; sorry Nelson for the logo) that would be able to handle the extra heat.
I then thought I would give a 15V 80VA toroid a try, followed by LM317 circuits (separate LM317 circuits for + and -, so 4 in total). It works, but the old hum monster was back.
And then I had a thought. The smps for my lottery vfet works fine and the smps for my b1 works fine, as does the 12VDC walwart I recently used for my H2V2. Why not try an smps? I have plenty of old 19VDC laptop smps psu's, but did not know if the output of these were floating or not, so did not try using one for each rail.
And then I came across a simple board on aliexpress, based on the TPS54360. It will take a single DC input and convert to a symmetrical output. Long story short, I fired it up today and it works. The best part is, no hum.
At the moment the configuration is (for each channel):
1 x 19VDC laptop smps
1 x TPS54360 board set to +/-15V
As the boards are relatively cheap (and may come in useful for other projects), I bought two. Since this is DIY and we are not afraid of overkill, I used 1 laptop smps + 1 tpss54360 for each channel
I have only been able to listen to it for a few minutes, but hope to get back to it in the coming weeks. I did notice some distortion when I had the dac at full (digital) volume, so really have to have a good look at which changes I made (if any) at the inputs in the past.
I have not had time to investigate the TPS54360 boards (besides check the output with a DMM), but it may be of interest to others.
The board I used is sold on aliexpress by a number of vendors for about US$7.
https://www.aliexpress.com/item/1005005815883872.html
Photo attached.
I bought a small enclosure from aliexpress (see photo; sorry Nelson for the logo) that would be able to handle the extra heat.
I then thought I would give a 15V 80VA toroid a try, followed by LM317 circuits (separate LM317 circuits for + and -, so 4 in total). It works, but the old hum monster was back.
And then I had a thought. The smps for my lottery vfet works fine and the smps for my b1 works fine, as does the 12VDC walwart I recently used for my H2V2. Why not try an smps? I have plenty of old 19VDC laptop smps psu's, but did not know if the output of these were floating or not, so did not try using one for each rail.
And then I came across a simple board on aliexpress, based on the TPS54360. It will take a single DC input and convert to a symmetrical output. Long story short, I fired it up today and it works. The best part is, no hum.
At the moment the configuration is (for each channel):
1 x 19VDC laptop smps
1 x TPS54360 board set to +/-15V
As the boards are relatively cheap (and may come in useful for other projects), I bought two. Since this is DIY and we are not afraid of overkill, I used 1 laptop smps + 1 tpss54360 for each channel
I have only been able to listen to it for a few minutes, but hope to get back to it in the coming weeks. I did notice some distortion when I had the dac at full (digital) volume, so really have to have a good look at which changes I made (if any) at the inputs in the past.
I have not had time to investigate the TPS54360 boards (besides check the output with a DMM), but it may be of interest to others.
The board I used is sold on aliexpress by a number of vendors for about US$7.
https://www.aliexpress.com/item/1005005815883872.html
Photo attached.
