The Multi Amp aka Dynalo Mk2

[Pars]

It shouldn't unless you decide you wanted to change voltages.

[cspirou]

5 minutes ago, Pars said:

It shouldn't unless you decide you wanted to change voltages.

Thats only for the positive regulator. The negative regulator uses a feedback resistor to determine the voltage.

[Pars]

Guess I should look at the neg regs datasheet.

If the pads are all connected together on the board, the solder bridges on those pins won't hurt anything.

[cspirou]

Success! After a few days of dealing with the regulators I finally got them soldered in properly. 

+V: 20.00V, -V: -20.02V

of course the 0.02V offset is super annoying even if it is within 0.1%. I calculated that I would need to add a 23.8 Mohm resistor in the AOT spot to adjust the voltage to -20.00V. Not really worth it, although I might just add it to my next Mouser order.

[MLA]

Nicely done!

About accuracy: it may also be that you have hit the limits of your DMM: 

 

[cspirou]

Thanks! Although I think its more likely the 10k feedback resistor is off by 5 ohms. Also I did a shitty job soldering that particular resistor. Redoing it might give me a better result. Or not.

[mypasswordis]

There is some variability between the internal reference voltages as well, but 0.1% overall is pretty damb chop good.

[spritzer]

Yeah, this isn't a laboratory grade voltage standard either.  If it were then everything would be suspect, how the board was constructed, internal stresses, are the solder joints contaminated and it goes on and on... 

[cspirou]

True, not to mention that it's not even up to it's running temp.

[Skooby]

Good job on the QFN soldering!  For some reasons, couldn't get the negative to work.  I reflow so many times that I thinked may burn thru the regulators.

[Pars]

On 2/5/2017 at 11:13 PM, Kerry said:

You really need to solder the large center pad or you could have thermal issues.  I don't think there's a way to do that with an iron.

If going with the onboard regulators, do these need heatsinks on the top of the board? If so, what, and how to fasten?

 

[mypasswordis]

I think I've posted this before, but maybe not. It's pretty easy to solder the qfn with a hot air rework station. This is the method I came up with, which hasn't yet let me down: Put the solder paste on, melt it and make sure every pin has exactly the right amount of solder. Adjust as necessary. Put some flux on and then put the qfn on, then blow some more hot air to solder it in place. You can gently nudge the part with tweezers to adjust if it's slightly off. Hopefully this helps people in the future

[luvdunhill]

Typically soldier paste takes much more hot air to flow the second time around... or it feels like it does.

[mypasswordis]

Hm, I didn't personally find that to be a problem but on boards where there is a large/thick enough ground pour connected to the pad then I could see why there could be an issue. 

[Pars]

I've never tried paste before but was going to order some on my next Mouser order (Chip Qwik 910-TS391AX Solder Paste No-Clean 15g Sn63/Pb37 T4).

I was asking about heatsinking the onboard regs (not offboard).

Also, for the offboard regs, any reason to use Amb's sig78/79 over some from ebay, such as seller LDOVR's (though I see he wants $220+ for the neg reg now

 

[cspirou]

Finished soldering my Dynalo Mini but I have problems. First thing when I switched it on the CCS LEDs did not illuminate. Checked voltage at the TRS jack and it was about 8-9VDC. The Meanwells work fine and show +24V/-24V.  The positive regulator is still showing +20.00V but the negative regulator is now showing +1.76V. This is very frustrating since it was measuring -20.02V before I started soldering the amp components.

Do I assume that the negative regulator is fried?

Edited January 1, 2018 by cspirou

[Pars]

I'd check for shorts on V-. Look for any possible solder bridges.

[ang728]

1 hour ago, cspirou said:

Finished soldering my Dynalo Mini but I have problems. First thing when I switched it on the CCS LEDs did not illuminate. Checked voltage at the TRS jack and it was about 8-9VDC. The Meanwells work fine and show +24V/-24V.  The positive regulator is still showing +20.00V but the negative regulator is now showing +1.76V. This is very frustrating since it was measuring -20.02V before I started soldering the amp components.

Do I assume that the negative regulator is fried?

Looks like some soldering issues to me,especially with tps7a regs

 

[mypasswordis]

8 hours ago, Pars said:

I've never tried paste before but was going to order some on my next Mouser order (Chip Qwik 910-TS391AX Solder Paste No-Clean 15g Sn63/Pb37 T4).

I was asking about heatsinking the onboard regs (not offboard).

Also, for the offboard regs, any reason to use Amb's sig78/79 over some from ebay, such as seller LDOVR's (though I see he wants $220+ for the neg reg now

 

Ooh, no refrigeration is nice. I have a separate compartment in my fridge for non foodstuffs, wonder how much it's seeped into my food over the years and made me insane. I put on the same heatsinks and thermal tape I used for the output devices on the regulators, and it does help some. 

[cspirou]

Here is what I get with an ohmmeter so far

Left vs right TRS - 79.6kohm

L/R TRS vs GND - 26.4kohm

+V vs -V rail - 4.67kohm

+V vs GND - 33.33 kohm (starts at 22Mohm)

-V vs GND - 5.835 kohm (30kohm when I reverse leads)

 

There was a couple resistors near the opamps that i fixed but didn't seem like an issue

 

i have a power supply i can connect to the holes for the offboard regulators. Is there any issue bypassing the meanwell and TSP7A and using my own supply for testing? Also could I just run the negative rail or do i need to run both rails?

[Pars]

There shouldn't be an issue using a bench supply/sigma22/GRLV, etc. for testing (don't plug the amp power in!). You would probably need to run both rails for anything meaningful.

[cspirou]

I have an lm317/lm337 bipolar supply I could use but it's not assembled. Just trying to see if I could get away with the single supply I have, but I'll just build it anyway.

[Pars]

You might be able to but I'm not sure what oddities you would see with any voltages you might want to look at. Maybe you could pull the IRM for the negative side and inject V- at the offboard reg point, powering the positive side with the internal?

[cspirou]

After a hiccup where a capacitor exploded on me because I installed it backwards, i finally have a functioning lm317/lm337 power supply. I adjusted it to +20/-20v with 10k resistors as loads and then hooked it up to the Dynalo mini through the offboard regulator holes, bypassing the onboard regulators.

After powering it on the rails dropped to +15v/-15v. At the TRS jack the voltage was about 0.4v instead of the 9v I measured before, but the output transistors were 220f! I also noticed the CCS LEDs weren't lightning up.

i didn't want to check other voltages while its running so hot. So ill turn it down to +9v/-9v before i start investigating further.

[Pars]

Ehhh, something sounds very wrong... the mini shouldn't load that supply that much, and it shouldn't be running that hot. Even with the PSU being plugged into the offboard reg holes, the onboard regs are still in circuit, with no easily removed component to isolate the amp channels from them (short of cutting traces). I would carefully go over the board, making sure that everything is in the right place and correctly oriented, and no solder bridges that you can see. You might also post some good pics of both sides of the board at sufficient resolution and maybe one of us might spot something.