SMPS minimum load

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

ruffrecords

Well-known member
Joined
Nov 10, 2006
Messages
16,152
Location
Norfolk - UK
I just spent a lot of time tracking down an annoying low level whistle. It is only noticeable at high gains (45dB or more)and with no signal (input shorted). The effect is to raise the noise floor by 10dB or so when measured with a typical noise meter. If you amplify the output by about 50dB and feed it to a speaker you can hear the whistle (sounds around 2KHz to me) but on the scope it is impossible to see in the noise (and I tried both analogue and digital scopes). I tried umpteen different version of PCB without any noticeable difference. I kept thinking it was a fundamental design flaw, especially as it quickly dropped in frequency as the power was turned off.

Then, after several days investigation I happened to attach the scope directly across the heaters and the analogue scope was able to lock onto the waveform - a mere 10mV pp. I tried adding a film cap but it made no difference. Thinking I had a duff heater SMPS I looked around for another to try in its place. Then a thought struck me. The SPMSU us rated for up to 8.5A of output current but I was drawing only 0.45A. Maybe the SMPS had a minimum current draw. So connected the power supplies to a rack into which I could plug several of the boards I had been trying. I fitted just the one I was currently testing, switched on and there was the whistle. Then I plugged in a second board and the whistle disappeared completely and the noise floor dropped to where it should be. I added a third board and it was still OK. I removed the second board just in case it was something special but still no whistle. Removing the third board so there was just the original brought back the whistle.

So a word of warning. It is necessary to over rate SMPS used for heaters so they can cope with cold heater inrush current but clearly at the other end of the spectrum they seem to like a minimum load below which they do funny low level stuff.

Cheers

Ian
 
This is known and some switcher designs deal with this in different ways. Some publish a minimum load, some skip pulses at low current. That might even be your audible (lower than normal) switching frequency noise.

Rather than over-sizing the switcher you could current limit it during cold start up providing a soft start until the heaters warm up... In fact some switchers have a soft (voltage) start built in. A capacitor on one pin defines start up ramp,  IIRC.

I inherited one switcher design that had a mechanical noise... This switcher provided two regulated output voltages from one IC, involving a trick transformer storage inductor. After exploring different loading I traced the noise down to the transformer's mechanical design and punted back to a simpler one output voltage topology using a silent off the shelf standard potted inductor, and added a linear pass regulator for the second lower voltage. Not as slick, but the customers didn't care about slick, just the odd extra noises.

JR   
 
JohnRoberts said:
This is known and some switcher designs deal with this in different ways. Some publish a minimum load, some skip pulses at low current. That might even be your audible (lower than normal) switching frequency noise.
I think this is highly likely what happened
Rather than over-sizing the switcher you could current limit it during cold start up providing a soft start until the heaters warm up... In fact some switchers have a soft (voltage) start built in. A capacitor on one pin defines start up ramp,  IIRC.

JR 
A very large number of SMPS have foldback current limiting which is totally useless for heaters - once rated current is exceeded the output current is limited to a small fraction of the rated current which is not sufficient to raise the heater temperature enough to drop their resistance enough to exit foldback limiting. I have even had this problem with a regulated linear supply (International Power) on an earlier project but fortunately I was able to defeat foldback current limiting and allow sufficient initial current to flow to warm the heaters.

A very small number of SMPS  will drive rated current into a short circuit but this generally means some very serious heat sinks and added cost.

Many modern SMPS now employ what is called a hiccup mode where they pass rated current into a short for a short period before restarting and trying to service the load. This is ideal for heaters because they get pulses of current of sufficient amplitude to quickly raise their temperature and drop their current demand. I have not yet determined what proportion of rated current one of these will successfully fire up into tube heaters so I currently conservatively rate them at 50%. They are so small and inexpensive I am not sure the added complexity of a soft start circuit is worth the effort. I have toyed with the idea of using a varistor but I have not looked at it in any detail.

Do you know of any simple reliable heater soft start circuits?

Cheers

Ian
 
Why do you care about inrush currents? Are you concerned that inrush current increases with load? I thought it didn't. But if it does, then maybe a choke on the output would help? I have always used a choke (but not for inrush currents - just to nix the high frequency ripple). It's not clear to me why you need to overrate the supply. If you use a constant current SMPS, it should supply the specified voltage up to the specified current only above which the voltage will start to drop.

And yes, if you don't load the SMPS it can throttle. This just happened to me as described in Re: Mini SMPS Mic Power Supply reply #3 half way down.
 
Would a series current limiter work?

IFD2634_F1.gif


Or an LM317 in current-limit mode?
 
squarewave said:
Why do you care about inrush currents?

I assume so it doesn't shut down during that moment.  One 8 channel preamp I had that drew 2.4A filament current metered over 5A at startup. 
 
EmRR said:
I assume so it doesn't shut down during that moment.
I must be missing something. My understanding of "inrush current" wrt SMPS is that when you apply AC the SMPS can draw a lot of current. For example, I have used MeanWell LPC-35-1050 which is 1A 30V. The datasheet reads:

INRUSH CURRENT(Typ.): COLD START 55A(twidth=500us measured at 50% Ipeak) at 230VAC

So 55 amps for a 1A supply. This is why I used this 10A switch:

nkk_s6a.jpg
 
JohnRoberts said:
Cold heater filaments are lower resistance until they heat up...
Ahh, I see. So it's "cold heater inrush current" and not SMPS inrush current you're talking about.

I would say just use a constant current SMPS so that it will naturally current limit until the heaters come up. But from glancing at MeanWell supply datasheets I don't see any 12V constant current SMPS with the right current range (1A-ish). APC-8-500 / APC-8E-500 is 16V 0.5A but the voltage is too high and the current is too low.
 
squarewave said:
Why do you care about inrush currents?
I don't but the SMPS does. Those with foldback current  limiting just will not fire up tube heaters.

It's not clear to me why you need to overrate the supply. If you use a constant current SMPS, it should supply the specified voltage up to the specified current only above which the voltage will start to drop.
I don't know of any constant current 12V SMPS. Can you point me to some examples?

Cheers

Ian
 
Yes, it is heater inrush current I am talking about not SMPS inrush current. Sorry for the confusion. My tube mixer modules typically consume 0.45A at 12V. So  mixer with 12 modules needs nearly 6A of heater current. Years ago I bought a big ammeter (10amps) to meausre heater inrush current and on a smaller mixer (6 modules)  with nominal 3A heater current, it would easily peak at over 6A when the tubes were cold. AC transformers for ac heaters and even linear dc supplies (without foldback limiting) will cope with this but many SMPS will not. I have found hiccup mode ones to be the best but I still over rate them to ensure reliable tube heat firing.

Cheers

Ian
 
It seems that simple current limiting should work with some safe margin above warm current draw, but fold back current limiting could latch low, as you speculated. 

JR
 
JohnRoberts said:
It seems that simple current limiting should work with some safe margin above warm current draw, but fold back current limiting could latch low, as you speculated. 

JR

That is right. As long as the current is limited to a significant fraction of the rated value then the heaters will warm up. It is only in recent years that SPMS that do this have become available. I guess when I get time I really should do some tests to find out the value of the significant fraction. All I know now is that it is at least 50%.

Cheers

Ian
 
mrclunk said:
What about NTC thermistors on each modules heaters?

That is a good idea. I had looked at a single NTC for the whole heater supply but I could not find anything the right value or dissipation. But per module is a different prospect. Thanks for the suggestion.

Cheers

Ian
 

Latest posts

Back
Top