|
Post by mkelly99 on May 7, 2016 22:46:37 GMT
I am not sure if this was the right way, but I'd like to feel out the reaction to me making a Raspberry Pi DAC Hat with the project ember architecture. I would change a few things, but fundamentally it would be the same. Of course, I would provide complete attribution to Frans de Gruijter/J.Helms, so that's why I am asking.
I have already done a 12V Hybrid DAC/AMP based on the Bravo V2, but wanted to go for one with higher voltage. I've looked at some of the John Broske Aikido stuff, but not sure I could fit four tubes! Looking at the Project Ember it's not clear what is allowed.
Thanks, Michael
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on May 8, 2016 12:16:26 GMT
Hi Michael,
I saw the Bravo alike design on SBAF.
What parts of the Ember design would you want to use ? I did design the Ember for Garage1217 and don't think it is good practice to just use that design. You can contact Jeremy about what specifics you want to use and which parts not.
I notified Jeremy about this thread, when he has some time I am sure he will react.
If you want you can use PM or e-mail for specifics in the design.
|
|
|
Post by mkelly99 on May 8, 2016 18:56:45 GMT
Most all of it. I might not have room for auto heater voltage select, and I might use a different buffer.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on May 8, 2016 19:55:31 GMT
Maybe it's an idea to use the RPi board to control the bias voltage. In that case you could make something that has similar features but goes about it in a different way.
All you need is 2 ADC inputs and 2 8 bit ADC to set the current source. There probably are in and outputs available for this.
Then you have something that is really different. When using another output device (so many to choose from) then it doesn't look anything like Ember (or Solstice) and you have something others don't have.
I design the G1217 amps so Jeremy can offer something different from what's already out there.
|
|
|
Post by mkelly99 on May 8, 2016 23:16:05 GMT
I will look at the idea of RPi based auto-bias. But, integrating it onto a Raspberry Pi HAT with the DAC already makes it rather different, no?
Michael
|
|
|
Post by mkelly99 on May 8, 2016 23:35:46 GMT
Also, would you be OK if I used your auto bias circuit for our 24V Hat? Our Kickstarter campaign is going really well and we could make this a stretch goal. All attribution to you of course!
Michael
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on May 9, 2016 5:02:54 GMT
Solstice has 24V based autobias.
When it runs on 24V it isn't an Ember derivative though but Solstice. What output device are you planning to use ?
|
|
|
Post by mkelly99 on May 9, 2016 9:46:12 GMT
Our 24V design uses the IRL510 and LM317T for the output stage.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on May 9, 2016 11:22:15 GMT
O.K. so it is not an Ember nor Solstice design but more like a Sunrise derivative. Except you want to make it autobias.
It is something G1217 doesn't offer but the thought had occured and a schematic with the idea was drawn. Will look for it ...
If you want to run your proposed schematics by me you can PM me.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on May 9, 2016 17:47:07 GMT
found it. This was a concept of the Sunrise with autobias. Wasn't built. What part would you want to use ?
|
|
|
Post by mkelly99 on May 9, 2016 21:33:24 GMT
Our design is virtually the same, but without the auto-bias. And we use SW to control the relay. One thing does make me hesitate on the auto bias. It might still be desirable to be able to run a higher bias for more linearity, especially driving low ohm headphone that require less swing.
Quick question on the schematic you sent. The gate resistor on the right channel is 470, but on the left it is 2.2k. Typo? Or is there some odd mismatching going on?
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on May 10, 2016 4:30:45 GMT
The schematic was just a thought process. In earlier designs 470 was used which later on was increased to 2k2. So values between 470 and 2k2 can be used. Both channels should be the same. (updated the schematic)
The schematic biasses itself if you want to change anode voltage you will have to change Z4 value. The question is if prefer asymmetrical clipping of a larger voltage swing over small signal levels with more linearity (lower THD values) for small signals. With a 24V voltage rail and ECC88 tubes linearity should be good.
When operating tubes on 12V supply voltages I suggest to look into ECC86 / 6GM8
As suggested you could also control it from the RPi perhaps.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on May 11, 2016 21:12:24 GMT
On your website I read:
The design is derived from the popular Bravo V2 architecture with all the modifications recommended such as: IRL510 Output Mosfet; Regulated Heater Voltage to reduce cross-talk; Adjustable Plate Bias for Tube Rollers; Nichicon MUSE Output Capacitors; and Jumper-able output impedance of 1.5, 32 and 100 ohm and switchable gain selection to match your headphones exactly.
The Bravo v2 doesn't have IRL510 MOSFETS, it doesn't have a regulated heater voltage, plate bias voltage adjustment is limited and on some Bravos not even present. It also doesn't have selectable gain nor 3 selectable output resistances (1.5, 32 and 100 Ohm). There is ONLY amp that has all these features and that is the G1217 project Sunrise.. Maybe it should be more fair to say the amplifier part is similar to the G1217 Project Sunrise instead of based on the popular (because its cheap) Bravo V2.
It is true though that the first Sunrise was based on a modified Indeed G2 though (not the Bravo V2 by the way).
|
|
|
Post by mkelly99 on May 11, 2016 21:36:12 GMT
I actually had not seen the Project Sunrise before we did this. I was working off the Bravo V2 and the various mods put forth on the forums. It might have saved me time if I'd seen Sunrise schematics first!
BTW, I gave garage1217 a shout out and a customer! Someone asked in the campaign comments for more clarity on how our hat tied into the Pi and they realized they wanted a stand alone amp. I put in the link to Project Sunrise III. Of course then they cancelled their pledge. :-(
Can I ask what your relationship with garage1217 is? Maybe you can work with us on future products? Although we are currently focused on the Pi, we are going to make broader market audio equipment eventually. I am the first to admit I am not a strong analog designer. Much more of a cookbook and implementation guy. I have the digital/computer side nailed so making products that combine them is something I can do very well. Michael
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on May 12, 2016 5:23:12 GMT
The 'well known mods' found on forums like RockGrotto (where many people link to) are mine.
IRL510 Output Mosfets; Regulated Heater Voltage (TSR2465), better Adjustable Plate Bias did come from my G2 mods. Jumper-able output impedance of 1.5, 32 and 100 ohm and and switchable gain selection are only found in G1217 amps (24V versions). For 48V versions its 1.5, 32 and 120 ohm. Not one of the Bravo/Indeed etc. amps have selectable output resistances.
I designed the schematics for all G1217 amps and assist with the PCB designs by giving comments/pointers to Jeremy. I am not an employee of G1217 but more like a personal friend of Jeremy and help with support and technical issues/problems.
|
|