[ARC5] Scope photos - Chinese DC-DC converters

[Ben Hall]

Hi Don and list,

On 10/17/2015 11:01 PM, don davis wrote:
> These converters should definitely NOT be putting out any 170/300 Hz.  There
> is nothing cornered that low in the loop for normal operation.  It might,

That is what I thought.  That shouldn't be there.

> without the start mode (soft start).  I assume the waveforms are with ac
> coupling and probes?

Yes - 100X high voltage probe, AC-coupled.

> The "hash" is the switching frequency.  In order to
> properly diagnose these you need a dc offset amplifier so you can see dc on
> top of the input dc without the common mode distortion, and a good dc
> current probe Tek TM xxx.  Look at the FET drain current w.r.t. FET Vds and
> look for cycle-by-cycle variation for shut down, LF oscillation, etc.

I'll have to look for both current probes and a DC offset amplifier. 
Last time I looked for current probes, the prices were pretty high.

> Other things:  What is the input supply?  Must not be current limited or
> high Z output - can get stuck in start mode.  Inrush current for start may
> be 10 times steady state!

This got me thinking.  The supply I had been using is a standard 
bench-top lab-type power supply.  In the past, I've had issues where it 
has done funky things with the current limiting, going into limit on a 
load spike.

So I got out a 12VDC gel-cell, 7AH, and tried it with some real short 
wires to the input of the supply.  Same behavior.  I was hoping it was 
my bench supply, as I'd really like to put these converters to good use.  :(

> Also no L in the input line - will oscillate at
> LF.  What is rated Vin and Pout?

These are the units I bought:  <http://r.ebay.com/MfTEUQ>

12 to 24VDC input, 200 to 450 VDC output, 40W continuous, 70W maximum.

> To get past obvious things have you tried
> running in the middle of the spec range for both?  Some dc-dc converters are
> just that - converters, and may not have proper input / output filters.
> This is common in acft (Interpoint, for example needs outboard filters to
> meet MIL-STD-461 CE01, CE02, CE06, and RE specs).

That is a great suggestion - just did a pretty quick test where I varied 
the input voltage keeping the load constant, and varying the load while 
I kept the input voltage constant.  While it was informal and I didn't 
do a matrix of the combinations, the 170Hz (on this unit) was there 
regardless of input/output conditions.  :(

It's interesting you bring up 461 and Interpoint.  We use Interpoint 
DC-DC converters in the aircraft electronics we build and my team used 
to be situated with a division that did EMI/EMC/EMV testing.  We have 
had good luck with the Interpoints, but they are pain to buy as they are 
not cheap and I've got to write a sole source each time we need more!

As you might can tell, I'm *not* one of the EE's on my team.  I'm 
actually an ME and now do more paper-pushing than anything - electronics 
/ ham radio is a hobby.  :)

> And, I did this for a living for 35 years in the aerospace business, where
> we have a technical name for designs that behave this was from the design:
> POFS

Hahahahahaha!  We use something similar.  ;)

thanks much and 73,
ben, kd5byb