The Multi Amp aka Dynalo Mk2

[Pars]

Ahh, that makes it clearer. The height issue is caused by the headers I am using (the plastic breakaway strip), so I can deal with that. I'll either get some nylon screws or drill and tap the heatsink for 4-40 or M3. The TO-220 silpads I have aren't quite long enough to cover the 3 annular rings, so I cut a strip and put it there so all is covered.

[cspirou]

So the dip-28 heatsinks are the perfect length but a bit too wide. One is wide enough to cover the row of pzta06/pzta56 thats in the middle.

Instead I'm using the heatsinks @mypasswordis mentioned from eBay. The heatsinks are quite light and the double sided thermal tape I'm using is pretty sticky so I'm not worried about anything falling off. I just hope dust doesn't accumulate in the gaps.

temp before was 75C and right now the heatsinks are 48C after an hour. Cool enough to touch for a short time 

[mypasswordis]

Did you set the bias lower than stock, and are you measuring like 30 minutes in? I think most people including mine are around 70-75 with heatsinks  mine would definitely go into nuclear melt down mode after being on awhile. 

I just sort of made a video for soldering the QFN chips for people here but ended up channeling some kind of cross between tangent and Patrick82 way too much while learning how impossible it is to solder with one hand and record with the other, so need to at least edit out the weirdest parts before posting (or remake the vid entirely). What's the best free video editing software?

[cspirou]

I'm using the stock 255 ohm resistors for bias. Although they are Panasonic instead of Vishay since they were in stock. I believe John said his reached 75C without heatsinks and that's matched what I measured after leaving it on for 2-3 hours. Its also winter time and the house is cold so I don't know how much of a factor that is.

However I did find that the that340 chip now measures 70C as Astro mentioned. I don't think that the heat is transferring efficiently to the heatsink and the gap between the transistors has made a sort of heat tunnel for the hot air to escape. The that340 chip is in between and is getting a blast of hot air from both sides.

I should redraw the vents for the GB enclosure 

Also Im measuring with an IR gun and it varies a lot around the heatsink fins. It would be more accurate if I attached a thermocouple to the heatsink.

[mypasswordis]

Oops, missed the part about measuring after an hour. Kerry and I confirmed in person that ours measure around the same temperature (both with heatsinks). Mine is run at 18V and biased correspondingly hotter from higher voltage LEDs, though. I used my multimeter with thermocouple adapter and measured right at the bottom of the fins. 

[luvdunhill]

Ahh, that makes it clearer. The height issue is caused by the headers I am using (the plastic breakaway strip), so I can deal with that. I'll either get some nylon screws or drill and tap the heatsink for 4-40 or M3. The TO-220 silpads I have aren't quite long enough to cover the 3 annular rings, so I cut a strip and put it there so all is covered.

There are M3 versions of those heat sinks. I wish I had some more. Someone here found them for me, so maybe a search would turn them up. I would definitely get a few extra of them if someone would send them over to us here in the US.

[cspirou]

When installing the heatsinks for the onboard regulators do you need any insulation? I just applied thermal paste to the heatsinks and then just placed them directly on the PCB. The thermal pad is tied to ground and the heatsinks are anodized so I wasn't worried about heatsinks having any sort of voltage potential. But if it's better to have a mica insulator then I can add one in.

[MASantos]

7 hours ago, luvdunhill said:

 

There are M3 versions of those heat sinks. I wish I had some more. Someone here found them for me, so maybe a search would turn them up. I would definitely get a few extra of them if someone would send them over to us here in the US.

 

If you send me the info regarding the right model I could help buying them here in europe and sending a package over the pond.

[UFN]

16 hours ago, luvdunhill said:

There are M3 versions of those heat sinks. I wish I had some more. Someone here found them for me, so maybe a search would turn them up. I would definitely get a few extra of them if someone would send them over to us here in the US.

 

I think that might have been me here? Fischer SK437 xx STS 2.

 

 

[luvdunhill]

I think that might have been me here? Fischer SK437 xx STS 2.
 
 

Yeah, those are they. Anyone else interested getting some of these to the US?

[Pars]

Sure! I’ll need to look at them however.


Sent from my iPhone using Tapatalk

[Pars]

Got the first of the 3 minis finished more or less (less case, etc.).

Had it running on AC for about 5 mins or so. 255R bias resistors, +/-20V PSU... I'm only seeing around 170mV across the 20R resistors. Low, and seems strange with others getting 18mA with 220R, etc.

As for adjustment (offset), I used the pots by the LEDs to get the offset between +/- for that channel centered around ground, then had dialed the individual offsets in using the other 2 pots. Anyone have a different adjustment method? I may play with the pot by the LEDs and see if that does anything to the bias current or not.

 

Edited February 2, 2018 by Pars

[cspirou]

I did the same thing for adjustment and it works well. I was able to get it within 3-5mV and the servos drop that <1mV.

I see discrete transistors instead of a that340. Nice

[Pars]

I did the same thing for adjustment and it works well. I was able to get it within 3-5mV and the servos drop that I see discrete transistors instead of a that340. Nice

Well the issue is the low bias current. 170mV is only 8.5mA. Target was 15mA. The bias resistors are 255R. I’m running 261R on my full size and getting north of 17mA on that (same PSU voltage).
After seeing the posts on here of people having to dial these back with 240R, 220R, etc. I was expecting higher. I’ll look at it some more.

And those are Linear LSK489/LSJ689 on some adapters I did.


Sent from my iPhone using Tapatalk

[cspirou]

Since you are using different input transistors I would think that has an effect. Are you using 422 ohms for the CCS? Should that be reduced back to 402 like it was originally?

[Pars]

The input transistors shouldn’t effect the bias current. My full size is also using JFETs (Toshiba).
I thought the 422R had to do with the LEDs having a higher Vf than the original 1.7V used in the full size versions...


Sent from my iPhone using Tapatalk

[cspirou]

I think Vf is close to the same. I measure 1.68V with mine

[Pars]

Sorry, it was in the group buy thread. Also he mentioned that R1 and R3 should be 422 ohms, not 402.
I got my Dynalo all back up and running, offset dialed in, GRLV. Listening, it sounds very good. I can't tell if it sounds better than it did with the sigma22; I think so but going from memory. I also had found that I had miswired the switch I use for SE->BAL input; it had left the - input floating when in SE mode, so I corrected that. Flipping back and forth, I couldn't tell any difference. Might have to actually listen to this with balanced in one of these days, if I can get my XLRs back from my son [emoji4]

I think Kerry’s original post was in the group buy thread (possibly deleted now?).
At any rate I quoted it in this post.


Sent from my iPhone using Tapatalk

[Pars]

On 12/23/2016 at 3:04 PM, Pars said:

On 10/6/2016 at 10:31 AM, Kerry said:

Here's the original BOM for the mini

SS Dynalo Mini BOM

There some mods.  The LEDs were changed out and the current resistors should be modified as well.  Also, there were a couple of changes for the resistors if you are using external regulators (AMB has the sig78/79 that can be used).  I think Joamat had some good values here.

The new LED is: LTST-C170CKT 

Also, the 402 resistors on the current sources should go to about 424 ohms.  This should bring the current sources back to 2.5mA.

Those are the differences I'm remembering.

^ Here is the original post on that.

Measuring mine, I'm seeing 1.64V across the LEDs (LTST-C170CKT). I believe 1.7V is the design spec?

1.06V across the 422Rs

Vbe for each of the  transistors by the LEDs seems low at 0.53V.

I let it run 45 minutes or longer. Bias voltage across 20Rs was up to 220mV, so 11mA. I think I will start by replacing the 422Rs with 402Rs. I don't really want to increase the bias resistors above 255R, but trying to balance sound with heat. This is in open air, and it is winter in Chicago, and in the basement. I was seeing device temps around 60C, a bit higher in spots (IR measured). I do have my full size amp downstairs to compare to as well.

Edited February 12, 2018 by Pars

[cspirou]

Be sure to use kapton tape to protect the Meanwell before doing rework. The plastic case on mine melted a little when I was removing LEDs.

[Pars]

Noted. If going to the 402Rs doesn't get me where I want to be, I may replace the LEDs with the Bivars that RudeWolf had in his BOM (1.9V). I had also forgotten to order the 0R resistors for the servos, so will solder on resistor leads until I get some.

I was disturbed to see how many items in my spreadsheet I forgot to order, even though I had built a separate tab which was a copy for importing into Mouser for the order. I went thru the order and removed backordered items, so that may be why (I forgot to find alternate source). Oh well

[Pars]

Hmm, must have forgotten to post the follow-on.

The 402R resistors got me pretty much where I wanted to be (262mV, or 13mA). Temps didn't seem to go up much when running it for half an hour or so.

On the first bank, I tried a THAT chip instead of the FETs and no effect on bias. Used hot air to remove the 2 over by the Meanwell and got them done without hitting that. Used an iron for the soldering.

So now I need to order more 402R as well.

Edited February 4, 2018 by Pars

[Pars]

Mini Dynalo

After running the amp for over an hour, bias was 14+mA, so right where I want it to be.

I thought it might be interesting to create a table or spreadsheet showing different users biasing and results, so I made a Google sheet, here:

https://docs.google.com/spreadsheets/d/1D5LDuRdupZRoWR-GfAd03fkFSBAMY4-E_VcYfCAI1Z0/edit?usp=sharing

Go ahead and fill in your info if you like so this is available for all to see. Note: this is the first time I've created a Google sheet, so let me know if you can't edit it and I'll try to figure it out

Heatsinks: I've been looking for some. The ones that @mypasswordis mentioned from ebay don't seem to be available any more, and I'm having a hard time finding any that appear to be suitable size. There are some I almost pulled the trigger on that were 50mm x 25mm, but the 25mm would leave the end transistors half covered. I could get the Fischer SK81/37.5/SA from TME and cut them in half, but I'm trying to avoid that if possible. I'd need enough for 3 amps.

EDIT: Found these, 45 x 35 x 18mm on ebay https://www.ebay.com/itm/263051780518?var=562054192200

Edited February 19, 2018 by Pars

[jose]

Doing a spreadsheet is a great way to share information about the behavior of the amp. Thanks Pars

[Pars]

Did a little "baking" yesterday, using the Kester EP256 paste I got. Still had to preheat the boards with a hot-air gun before applying the paste, but I like this paste a lot better than the Chip Kwik I had gotten. I think it turned out pretty well.