The Multi Amp aka Dynalo Mk2

[Whitigir]

I have not,  but i think they are sitting 110-120C.  Can anyone else confirm about this at 90 minutes?

Edited November 4, 2018 by Whitigir

[Pars]

Without the heatsinks, I am seeing the following (IR gun):

~45 minutes (no music playing)

Top: 41C

Transistors (hottest): 102C

Bias is a bit hot at ~345-360mV = 17-18mA (this is with 402R resistors and the LiteOn LTST-C170CKT LEDs (1.56V drop).

Maybe I'll toss in some 422R for the bias and recheck; that had always measured low initially (12-13mA), but I'd never run it this long in that configuration.

[Whitigir]

Thanks @Pars.  To account for the IR gun differences, I think it lines up with what I am reading.  These things can be cooking .  Considering it performances, little form, power output.  It is amazing though

Edited November 4, 2018 by Whitigir

[Pars]

I just ran pretty much the same on the 2nd amp, with more or less the same results. I'm going to play around with the bias a bit and will report back, though these things normally sound the best biased as high as you can get away with (within reason; I think Kevin had commented regarding the full-size that around 20mA was the sane max).

[jose]

102ºc is not a bit overkill? 

 

[RudeWolf]

Well, solder stays solid and transistors themselves are rated for temps like these. I just don't like the idea of thermal expansion putting mechanical strain on stuff. Of course adjacent parts can age faster due to thermal radiation.

[jose]

I ask from ignorance. I have the Dynalo MkII "normal" (TH components with SMD sand) and if I remember correctly, transistors were 40~45ºc and I already seemed very hot for to be SMD.

I think the maximum temperature of the sand is about 150ºc but I wonder if this will not be bad for the other components or the pcb.

Edited November 4, 2018 by jose

[Whitigir]

The other components actually will not be as hot as the transistors.  However, the heat being dissipated around the heatsink May heat up the adjacent components by a bit but wouldn’t be anything like 100C.  

[Pars]

Rechecking with reduced bias (13-14mA). Max device temp (no heatsinks) that I measured was 92C (R- bank, back around the AC/DC SMPS). Most are around 80-85C. Still sounds good, maybe not quite as good, but not really noticing much difference sound-wise.

These temps are just an artifact of the form factor. Unlike your full-size, the output devices on these are concentrated in a 4"x2.5" area. My thru hole full size runs much cooler as well.

[MASantos]

There is space inside for one of those ultra quiet computer fans. With a reduced voltage removing  a bit of air from inside the box might help.

[gepardcv]

Wait, did anyone actually determine that the heatsinks don't help cool the transistors? That's... counterintuitive.

[cspirou]

1 hour ago, gepardcv said:

Wait, did anyone actually determine that the heatsinks don't help cool the transistors? That's... counterintuitive.

The heatsinks attach to the plastic casing which I presume isn't a great heat conductor. Not too surprising. Doug is using a heatbar attached to the PCB for his T4 amp, which is a pretty clever solution.

Also wondering if this is why my amp with the 3oz board has a lower temp. More copper to dissipate heat.

[Pars]

How much lower temp is it? Also, what were your impressions regarding building ease of the 2oz boards vs. the 3oz?


Sent from my iPhone using Tapatalk

[cspirou]

I haven't built a 2oz board yet. Just comparing the reported temp of @Whitigir vs what I have, which is closer to 70C max. I'm sure building the 2oz board is much easier. Although the 3oz board wasn't so bad after I got a soldering strategy down. Plus I have a DIY reflow oven which helps with a lot of components.

[Pars]

Out of curiosity, what did you use for LEDs and the CCS resistors? 402R or 422R?
Note that the temps I am seeing(IR gun, no sinks) are higher than what you show, even at lower bias.


Sent from my iPhone using Tapatalk

[Whitigir]

These seems to be a possible fit, if u do use paste, and maybe doing a better job ?  Now, I wonder how do I take off the sticky double sides thermal glued to remove my heatsink....it sticks so so strong 

https://www.mouser.com/ProductDetail/Advanced-Thermal-Solutions/ATS-51450R-C1-R0?qs=sGAEpiMZZMttgyDkZ5WiulqfIpz7MfcldLiKLTRn2uU%3d

Edited November 5, 2018 by Whitigir

[Pars]

You have to carefully pry the heatsinks off. They will come off without damaging anything, just take your time and be careful about where you are prying.

Those heatsinks you linked are rather large, not sure they would fit, particularly with FET inputs (TO-71).

Any comments on the biasing questions I asked above?

[Whitigir]

2 hours ago, Pars said:

You have to carefully pry the heatsinks off. They will come off without damaging anything, just take your time and be careful about where you are prying.

Those heatsinks you linked are rather large, not sure they would fit, particularly with FET inputs (TO-71).

Any comments on the biasing questions I asked above?

Ah, I thought it was for @cspirou, I am using C170CKT and 402R

[rumina]

Another smd Dynalo is born, verry happy that i could manage that. First time hand soldering so much smd parts und my eyes getting older. Was not easy but worked out and the sound is great. Gets very hot with the 422 402 ohm, same heat as the one from pars.

I use my seekpro form some thermal pictures after 45min, for me ok when i use it but wouldn't give it away with this heat. For the Heatsinks i used thermal adhesive, due the different ems from raw alluminium the thermal camera show the temp to low, they are hotter then the pcb but not over ca 80 c. Hottest parts are the small transistors.

Tip to measure raw alluminium with a heat gun stick a black adhesive tape over it.

Thanks for the group buy of the pcb, the group buy of the case and all the support i get from this great forum.

 

 

Edited February 2, 2020 by rumina
typo / correction

[Pars]

Nice! So you’re using 422R and the LTST-C170 LEDs? 255R bias resistors?
What are you getting for V drop across the LEDs initially and after 30 mins or so? And what bias on the 20R output resistors (V drop / 20)?

If everyone building these could post this info, it would be appreciated! Maybe I’ll toss a google sheet up to record this so everyone can access.


Sent from my iPhone using Tapatalk

[rumina]

Sory pars, i used 402 ohm, had it wrong in my head.

I use Bivar Leds SM0805HCL, V drop after 45 min is 1.705 - 1.715 V This with 402 ohm and 240 ohm (instead 255 ohm on the pcb).  Can meausre the bias over the 20r, it's buried under the glued heatsink. 

bom i used is the one from kerry - thanks btw for this :-).

[Whitigir]

You used perms-epoxy @rumina ? For heatsink on Transistors ?

[rumina]

As i understand it's a mix of thermal paste and some epoxy, i used it often for ntc putting on transistors. https://en.wikipedia.org/wiki/Thermal_adhesive

It's not as good as only thermal paste but better then nothing. The trick is to use not to much. I think the performance is better then silicon thermal pads - but it's then glued - no easy fix afterward.

Edited November 6, 2018 by rumina
typo

[Pars]

Understood... thanks for the data. Those are the other LEDs I have. I was thinking about trying the 402R with current C170 LEDs ( which I am getting 17- 18mA bias) and dropping the 255R down to 240R and see what I get when warm.

 

One thing I don’t quite understand about the operation of these CCSes is the increase I see in bias of around 5mA or more between cold and hot. Is this caused by the LEDs being heated? Doesn’t seem very “constant” to me...

 

EDIT: Put 240R in one channel. After warming up 30 minutes, it was right where I wanted it (290-304mV or 14.5-15mA). The other channel was at 340-360mV, or 17-18mA. This is 402R and the LSTS-C170 LEDs. My one concern is at startup, bias is only in the 8-9mA range, which from experience sounds a bit flat.

EDIT2: And to my last previous question, I guess the output transistors heating up probably have more to do with the bias increasing with heat than the LEDs do.

 

 

Edited November 7, 2018 by Pars

[rumina]

Have you measured the power consumption of your smd dynalo?

After a amps is "near" finished i like to measure the power consumption over 90 min.

On the one side you can estimate the real class a output and whats even better your can see when the amp get into a thermal stable state. Normaly, as the amps heats up, the bias and so the power consumption rise up to beeing stable (or a atomic meltdown ? ). And yes, it's the easy way to guess the output, missed that part in the school...or im to lazy....

my smd dynalo starts at 27 watt and gets stable at around 30 watt, near the max wattage of the two trafos. But its nice that the trafos are protected, i think bevor the amps overheats/runs away the trafos shut down at 115% (~35 watt). So no need for a thermal saftey switch i used in some pass amps i made.

would be nice to see how much energy the other smd dynalo here are burning.