The Multi Amp aka Dynalo Mk2

Aive · September 3, 2014

Those LEDs you listed are 2.2 v forward voltage drop.... I think you need the 1.7 V variant if it's anything like the dynahi?

GrindingThud · September 3, 2014

They drop 1.7 in use for me....and that is what is listed on the data sheet as typical.

Those LEDs you listed are 2.2 v forward voltage drop.... I think you need the 1.7 V variant if it's anything like the dynahi?

Edited September 3, 2014 by GrindingThud

Aive · September 3, 2014

Ahh ok, I only checked the Mouser spec page.

*spritzer* · September 3, 2014

This is what we use for 1.7V led: 859-LTL-307ELC

the lead spacing on the board for the .1uf cap is 3.5mm but the .1inch caps do fit.

the 5pf caps are 5mm lead spacing

by the way, I have a new name for the thing...

PureBipolar (tm&c)

would apply to the ss dynahi too.

Also the footprint for the resistors on this one is 0.5 and not 0.6 as on most of the other designs.

Edited September 3, 2014 by spritzer

*kevin gilmore* · September 3, 2014

that is because (among other reasons) birgir wants things to be as tiny as possible.

*spritzer* · September 3, 2014

It offsets my petite stature...

nopants · September 4, 2014

any downside to running at 15V? I thought I spied 20 on the silkscreen

GrindingThud · September 4, 2014

Mmm, parts incoming.

*spritzer* · September 4, 2014

any downside to running at 15V? I thought I spied 20 on the silkscreen

Less voltage swing and power aside, there are none.

*mypasswordis* · September 7, 2014

any downside to running at 15V? I thought I spied 20 on the silkscreen

The OP07 datasheet specifies a maximum of +-18V supply voltage so I personally wouldn't go above if using that opamp. I am going to run mine at +-16V (mainly based on the 15VAC transformers I have on hand).

*spritzer* · September 7, 2014

OPA445 for the higher rails.

GrindingThud · September 7, 2014

What is the preferred sequence for adjusting the pots on this unit? The 3rd one is messing with my mind.....

*kevin gilmore* · September 7, 2014

remove the opamps. make sure the 2 pots that adjust output voltages are in exact center

hook it up, and adjust the led bias pot for best output voltages around zero.

put the opamps in, and adjust the other pots for pin6 on the opamp close to zero.

GrindingThud · September 7, 2014

Got it, thanks again!

*luvdunhill* · September 7, 2014

I have a bunch of matched jfets I was planning to list, but life is too crazy now... I drug them out of a pile and moved them to yet another pile and am not sure where they are, or I would check. Hit me up week after next and I should be able to have time to locate them.

I can supply LSK170 in 1% matched quads (or higher). I don't really want to part with the p-ch part though. Anyone think there would be any demand for these?

*mypasswordis* · September 7, 2014

Thanks anyway Marc. I have 2SK170s but don't have any p channels.

*mypasswordis* · September 12, 2014

I did some board stuffing while I think of what to do about the j74s. I read that the Linear parts are different enough from the Toshibas that they are not direct drop-ins so I can't do complementary pairs from different manufacturers. Decided not to bother with offboarding the CCS offset adjustment so I'm going to just put in the 20k resistors. Need to put in another mouser order and populate the attenuator boards, so I can actually start programming the damn thing. Preliminary testing is good with the uC, it connected to my phone through the cloud just fine with the wifi antenna I bought. It's a small planar antenna with an adhesive backing so should be easy to mount to an enclosure.

Just want to make sure I have the ubal to bal boards traced correctly so I don't build it wrong (not sure if a schematic has been posted). It's a cascoded diff pair with source degeneration and a CCS tail into a wilson current mirror driving output BJTs? Are the LEDs across what I think is the mirror used for Vbe temp tracking or some other purpose?

*kevin gilmore* · September 12, 2014

the boards you have are the ss dynalo boards, here is the schematic

http://gilmore.chem.northwestern.edu/kgdynalobalssproduction.pdf

your boards are slightly different. work fine with the that340 part

and a 10k pot. just amazing sound into all headphones I have tried.

the unbalanced to balanced thing is

http://gilmore.chem.northwestern.edu/ubaltobald.pdf

and really the that340 needs to be replaced with fets for high impedance

sources

*mypasswordis* · September 12, 2014

Thanks! Yes, my boards are missing the CCS adjustment but they have the trimpot for offset on the servos. I saw the Dynalo SS schematic previously but didn't see the ubal to bal one.

I noticed the input diff pairs aren't cascoded in your schematic but seem to be on the board? Also, would you recommend putting in a CCS adjustment for the ubal to bal CCSes?

*kevin gilmore* · September 12, 2014

the pots on the opamps will actually do very little. and the front end is not cascaded

the 10uf servo caps need to be 1uf and the bias resistor should be 240 to 255 ohms.

change parts as necessary to make that board identical to the schematic.

Edited September 12, 2014 by kevin gilmore

*mypasswordis* · September 13, 2014

the pots on the opamps will actually do very little. and the front end is not cascaded

the 10uf servo caps need to be 1uf and the bias resistor should be 240 to 255 ohms.

change parts as necessary to make that board identical to the schematic.

Yeah I'm sure it does very little as the OP07/OP27 max offset is already 1.5mV and the thermal stability of trimpots aren't very good. I should have followed the datasheet with a 20k pot but got 100k instead. For anyone that has these older version boards and want to put in the trimpot, you're going to also need to jumper the NC pin to the offset pin if using the OP07/OP27 since the pinout of the OP07/OP27 are slightly different from the OPA445.

The servo caps I put in are 2.2uF MLCC, will use 255 ohms bias resistors and will add in the LED bias adjustment after all since you said so. The bias resistor slots are actually marked 510 ohms on the silkscreen. I have followed the values in the schematic over the silkscreen for the other resistors as well.

So I should also follow the schematic and not the board for the ubal to bal as well?

The CCS resistor value is 100 versus 180, there's 50 ohms in series on the output, and again I believe the front end to be cascoded with the extra BJT and the voltage divider from the rails to gnd connected to the bases but could be wrong.

Going to make this as KG approved as possible eventually since it's the KGITSOCH, I will build scaled down Dynahi power supplies for this thing. It will have three transformers with pass FETs connected to offboarded heatsinks to give the feel of the DIY T2 supply. I will also put in the MM stage you posted somewhere that you designed in the '70s unless you have a newer design.

Edited September 13, 2014 by mypasswordis