Group Build: Dynafet

[Pars]

The Dynalo servo was supposed to be able to handle 50mV, but in my experience, even asking it to handle 20mV is a problem (though 20mV really isn't a problem for most headphones anyhow.) I prefer to get offset well inside of 5mV or even +/-2mV before throwing the servo opamp in, but that is probably just me being anal. In the original dynalo, offset was dealt with via LED substitution and resistor selection for the CCS's. Justin's V2 (and I presume V1) boards also did offset reduction this way... not sure what tolerances Justin did his production boards to or how he even handled this. This was rectified on the Dynahi boards by including the trimpots paralleling the CCS resistors.

I know that it would take the servo awhile to do this (as you stated in your posts with digger), probably on the order of 5 minutes to stabilize. It would seem that selection of parts (R55, maybe R19/R20) could balance this out, but maybe not.

BTW, I found and am using a single pin connector for the outputs on this (and the dynahi) which might be useful to GB members. I can't find an equivalent at Mouser though. These require drilling a hole in the PCB, but that isn't a problem. And they will fit even with all output devices populated on the DynaFET. These are JST XH series:

header: 455-2227-ND CONN HEADER XH TOP 1POS 2.5MM

housing: 455-2220-ND CONN HOUSING 2.5MM XH 1POS

pins: 455-1135-1-ND CONN TERM CRIMP XH 22-28AWG

Edited February 11, 2011 by Pars

[luvdunhill]

The Dynalo servo was supposed to be able to handle 50mV, but in my experience, even asking it to handle 20mV is a problem (though 20mV really isn't a problem for most headphones anyhow.)

which opamp was used, how large were the voltage rails, and were local regulators used? that seems crazy low.

[Pars]

I usually used TL071s on 15V rails, like all Dynalos.

[luvdunhill]

yeah, so the opa445 with 30v rails will have considerably more range in it.

[Pars]

I'm sure it does, and that is the reason for an opamp that can run from 30V rails. I still prefer to get offset as low as possible before relying on the servo for clean-up. And 160mV doesn't seem low to me. I can adjust it out via the pots, but there wouldn't be any quick switch between the two modes. I know Kevin told me you would need the servo for the local feedback option, so maybe there isn't any way to balance things out for low offset (without servo) for either feedback method simultaneously.

[luvdunhill]

I'm sure it does, and that is the reason for an opamp that can run from 30V rails. I still prefer to get offset as low as possible before relying on the servo for clean-up. And 160mV doesn't seem low to me. I can adjust it out via the pots, but there wouldn't be any quick switch between the two modes. I know Kevin told me you would need the servo for the local feedback option, so maybe there isn't any way to balance things out for low offset (without servo) for either feedback method simultaneously.

what about split the difference at 80mV?

[Pars]

^ probably the best compromise if you want to be able to run either way and there isn't a more elegant solution. Once I do some listening to it, I would probably decide I liked it one way or the other anyway and tweak it for that. End of problem.

The FET instability is more worrisome however. I've kinda put this down for awhile, but am interested to see what you encounter when you get a board (or more) built up and start looking at it.

[luvdunhill]

yeah, should be somewhat soon... I have a part order to get ready (it has several long lead items, also a few sole source items like some transformers from Scientific Conversion and oscillators from Tent) then I'll get back to this, hopefully this weekend. The resistors arrived from PcX and they measure properly. Yay!

[Pars]

Gotcha. PM incoming...

[DigiPete]

Anybody contact J4CBO, surely he has the eagle files since he designed the boards

are these it?:

Edited February 23, 2011 by DigiPete

[Pars]

I think those are the prototype through-hole files that I have. They don't match the gb board exactly. I tried contacting Jacob, but no response. I used the contact link through HC... not sure if that is a PM or email. If anyone has his email address, we could use the eagle files for both versions. Any other documentation he might have (BOM for the SMD version, etc.) would be gravy. I asked Kevin and he doesn't have them.

[DigiPete]

maybe try him on Head-Fi or HeadWize?

[vvs_75]

Just curious if there is anything new to report?

[Pars]

I put mine away back in Feb. and haven't touched it since. I was waiting to see if Marc got his going or had some input but haven't seen or heard anything since then.

[Pars]

Anything further from anyone? I haven't heard of anyone with the lil'knight GB boards discussing building those either. I may take another look at my previous GB boards (didn't buy any of the newer ones).

Edited February 13, 2012 by Pars

[naamanf]

I'm also interest in this. Would like to think I all those parts are not going to go to waste.

[Dreadhead]

Any news? I'm not building but my buddy and I have quite a bit sunk into these....

[Pars]

Haven't looked at mine...

Seems like j4cbo is unloading all of his stuff (see FS threads), so looks like he never got his built. He also hasn't weighed in on any of the threads with any build/turn-up experiences (that I have seen... may have been a couple of posts over on hf).

Edited March 20, 2012 by Pars

[deepak]

Has anyone that bought the new boards from Lil Knight got theirs up and running? I recall shipping out quite a few of them to people.

[DouglasQuaid]

I haven't started populating any of the lil knight boards, even though i bought sets for both Dynas and a KGSSHV. Also slightly regretful I'm late to this thread, as I was planning on putting a dynafet together sometime in the next year or five.

[Pars]

It was brought up in the GB thread that few (if any) had gotten one running...

Edited March 22, 2012 by Pars

[kevin gilmore]

I certainly have not built one yet. Don't have enough time.

[Pars]

Anyone have the latest DynaFET schematic, preferably showing the feedback options? The only one I am finding is the .png one that j4cbo had...

 

Nevermind... I found the balanced DynaFet schematic so can discern what I might be looking for from that.

Edited February 27, 2015 by Pars

[Pars]

Being a glutton for punishment (and since I was messing with my dynahi anyhow), I decided to take another look at this. Nothing yet, other than a few observations:

1. I built an LTSpice model for this (attached, change extension to .zip), which seems to work. I think I found good enough models for devices.
2. In DC testing, some of the BJTs in the DB sections are running rather hot (55C). Q19/Q22 and Q23/Q26. Bias for 4 output devices is ~80mA per device.
3. Current configuration of the Vbe multiplier is 2.2K for R17/R18, 1K for R62 (not on Eagle schematic but in series with TR3), and TR3 = 20K. Adjustment range of ~88mV or so low to > 150mV. Current source resistors are 1 ohm but probably going back to 2 ohm. I currently have 1K gate stoppers in, but have not tried running signal thru yet to look for oscillation.
4. Eagle files as I have them are available on request. Not completely correct, but close to board.

I'd like to have this thread moved to the diy section, unless someone objects. Hard to find here.

 

 

Dynafet ltspice.doc

DynaFET_schematic.pdf

[Pars]

Promising... actually listened to single channel on some crap phones tonight. My scope isn't showing anything bad (though I think my scope may need recapping). More testing, etc. necessary. One thing I do have is some hum/buzz; this increases with volume until the pot gets past 90%, then decreases to nothing. No buzz all the way down either. Not sure if this indicates noise/oscillation or not, but what was on the scope wasn't conclusive.

One other change not mentioned above was increasing the feedback resistors from 1K/100R to 100K/10K (R56/55). I noticed that on the schematic for the SuSy DynaFET that feedback was shown as 200K/10K; I assume this is mostly to increase input impedance, but figured going higher might not hurt. I'll switch it back and see what effect that has.