• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Output current instability in DCDC module...with higher C(out)

Flyback

Well-Known Member
Hi,
We find that when we have more than 220uF of C(out) on our DCDC module the output current gets unstable as attached.
It’s a DCM3623T50M53C2T00.
The Eval board for this module only has 220uF.

Our spec is vin = 34v, vout = 48v, pout = 312w
Transients are 0-100% load

However, when we add say 470uF or 2mF or more, the current is unstable, but the instability is always bounded and never gets worse than the attached, no matter how much extra capacitance is added.
The vout ripple remains fine.
Its still able to supply max load no problem.
Do you think it’ll be OK to use it with >220uF?
We cannot adjust the feedback loop components, its internal.

And what about having 40 of these in parallel, each with the same instability?

(the current in the attached is (v-0.33)/0.264 Amps
(its the output of a ACS722 current monitor)

.......*.................*........................
.I must admit this noise in the output current is the most unthreatening noise i have ever seen...it doesn't get any worse no matter how much extra c(out) is added.....it is always bounded to +8% to -8% of average...and the vout ripple is fine...and the load service is fine even when 0-100% transients.....the noise involves smooth changes in current. It really is the mildest noise ever.
Do you think it could get worse under some condition?...we cannot change the feedback compensation as its inside the module internally. Looks like we're stuck with this noise if we want c(out) > 220uF.....but then again...it hardly seems a problem. The DCM3623T50M53C2T00 datasheet says if we want to be able to serve transients down to <10% load then we need >220uF for c(out).
.....................*..........................*............................*............................
The interesting point is that we also have an external error amplifier to control the TRIM pin of the dcdc module, and interestingly, when that is functioning, it gives a perfectly stable TRIM pin voltage, and successfully controls the output current to the correct average value…… but yet still the noise occurs on the output current………so in that case, due to the stable TRIM pin voltage, it cannot be said to be a feedback loop problem. I think this noise is some kind of thing going on in the internals of the module. Would you agree?
............................*............................*.....................*......................*
This “noisy output current feature”….is possibly no bad thing….the module still supplies the load perfectly well with max/off transients served very well.

Also, I have just put 12mF of C(out) on the output, and the output current “noise” is no worse than with 470uF on the output…so I am changing my view, and saying now that this “noise” is actually possibly a special feature of the vicor DCM3623T50M53C2T00 module, and is part of the scheme of it being able to handle very wide ranging amounts of C(out)…would you agree?
 

Attachments

Last edited:

Flyback

Well-Known Member
The following scope shots show the output current “noise” on various timescales…as you can see, it does not look like noise, and looks like some kind of active regulation scheme going on inside the DCDC module……these shots were with the external error amp doing the regulation , and that feedback loop was perfectly stable with a smooth , stable TRIM pin voltage…so there is no unstable feedback loop here.

(The noise is the same when I have the module on static TRIM).

These shots were taken with vin=31v, vout = 40v, iout = 3.5A (external error amp)

Also, its notable that the “noise” looks very similar to the Burst mode output current pulsing that goes on when the module is on no load…..

The "noise" does not happen when c(out) <220uF.
 

Attachments

Latest threads

EE World Online Articles

Loading

 
Top