Toroidal inrush

[McMurdo]

Ok!
Using my inline monitor, I have 0.09A, 9.9W, power factor 0.44

[mhennessy]

Originally Posted by Craig Sawyers

Looking at the power draw and thermal for this beast from https://www.crownaudio.com/en/products/xli-3500

The maximum current draw from the mains when delivering 660+660W is 22.5A

Only it won't be. It's designed for live music. The peak to mean ratio here will be massive, so the average draw will be much less than the peaks. Any attempt to use this amp to drive a sine wave at those quoted powers into a dummy load would cause an over-temperature shutdown. But then, I know you know that

Not sure where 660W comes from. The power specs are a bit ambiguous on the website, but I think they mean 1000 watts per channel into 8 ohms. This is "continuous average sine wave power" in theory, but only for short bursts or much higher crest factors in practice. It would be better to say "90 V RMS", IMHO. Into 4 ohms, this only rises to 1350W per channel, which shows that the linear PSU is not very stiff. When bridged into 8 ohms, you get 2700W into 8 ohms, which is a long way from the 4kW you'd get in theory, but consistent with what you'd expect given the 4 ohm value.

So maximum power output ranges from 2kW to 2.7kW. Assuming a power amp efficiency of 70%, that's a power requirement of 2.8 to 3.8kW from the PSU. Ignoring its losses, that's 16.5A at 230V.

Of course, the IEC input connector is only rated at 10A. Obviously it'll handle short-term overloads, providing the average stays below 10A, but I think the limiting factor here will be the mains transformer. We don't know the physical dimensions, but if we did, we could estimate its VA rating - I wouldn't be surprised if it was only 500VA - certainly no more than 1000VA, I'd guess (but those are guesses based on experience with these sorts of amps). Zooming in on the graphic in the user manual, the IEC inlet is labelled "700W". That's consistent with the likely VA rating of the transformer.

The rear panel in the manual also says "220V". Earlier I suggested that it's worth finding out the point where the magnetisation current shoots up because of saturation. Having spotted that, I'd definitely recommend doing that now! If your local mains is 245V - as mine sometimes is - then that could well be part of the problem.

Obviously we don't know how well designed that transformer is, but a regular "off the shelf" 500VA transformer with a primary designed for 230V will be grumbling rather loudly with 230V mains because of saturation. For hi-fi, you need to run at a lower flux density to avoid that, which is why the 400VA transformer I mentioned earlier is the same physical dimensions as a standard 500VA transformer - in effect, it saturates at a higher input voltage, so runs silently with the normal range of mains voltages you'd expect. But PA amplifiers aren't installed in quiet listening rooms, so silent operation is not a requirement (there's fans as well) - that means costs can be saved in the transformer design...

[mhennessy]

Originally Posted by McMurdo

Ok!
Using my inline monitor, I have 0.09A, 9.9W, power factor 0.44

That's a bit on the high site, but not outrageous. What was the input voltage? And using the variac, is it much less at 220V?

[McMurdo]

I should just mention that it's perfectly possible to run this amp with a series 100W lamp limiter, in which case it starts up and once enabled/unmuted, does not light the lamp. There are two inrush blips as I already saw on my bench ammeter: one at power-on and another when the bypass relay closes. I'll try to film it before disconnecting the secondaries again.

[McMurdo]

PS..the toroidal data label is marked pri: 240V 50Hz.
With the amp powered up fully (no audio) it takes 150mA @ 220V and 200mA @ 245V

[Radio Wrangler]

Those figures aren't at all outlandish.

I think there are two big transient effects, one is the charge-up of the capacitor banks from absolutely zero. For that sort of amplifier, I think this is intended to be a rather large energy store, plenty of Farads run to quite large voltage... the voltage nbeing needed for those large transient powers on the outputs.

secondly, as already said the toroids themselves can give large start-up transients depending where on the cycle the switch (or relay) closes.

Transformer designers and their accountants in the back office do get carried away when toroids are mentioned. GOSS can be used efficiently which gives the possibility of running to rather high flux densities and using fewer turns per volt. THis saves money, but makes for transformers with hellish turn-on transients, and also Frankenstein film set levels of audio sound effects.

My big power amp makes no claims to those power outputs, but it does cause a flicker in the lights when started-up. The trips stay in. My big tool transformer (think it's rated at 7kVA) is about equivalent in flicker.

David

[mhennessy]

Hmm - between 33 and 49VA at idle doesn't seem too bad, really. The raw magnetising currents are higher than ideal, but not indicative of shorted turns or other problems IMHO.

Good that the transformer itself is properly spec'd in terms of primary voltage. I wonder if your example has "220V" printed by the IEC inlet? Not that it matters, just curious...

Most likely your MCB is just a bit too quick. Do you know what speed rating it is? There should a letter on there to indicate this - usually B, C or D.

Either way, if the amplifier's rear panel breaker is holding up, then I suspect all is well. You might just be "lucky" to have an especially low mains impedance. Try powering the amp via a 20m+ mains extension cable - that might do the trick

Is the relay unpluggable? If so unplug, wait a while then plug in.

[broadgage]

These transformers can have very surprising inrush currents.
A crude test is to power up the amplifier from the most distant mains socket via a couple of long heavy duty extension leads.
The slight extra resistance introduced by say 50m of 1.5mm flex, plus the extra resistance in the extra length fixed wiring caused by use of a relatively distant socket outlet, may avoid the blowing of fuses.
Or try powering up in series with a kettle or space heater, and bypassing this after a second or two, similar to a lamp limiter but with a 2or 3KW load rather than a lamp.

If the amplifier then works correctly, that suggests a simple inrush current problem and not any underlying problem.

Review the series NTC device and ensure that this is not bypassed too promptly.

[Craig Sawyers]

Originally Posted by mhennessy

Only it won't be. It's designed for live music. The peak to mean ratio here will be massive, so the average draw will be much less than the peaks. Any attempt to use this amp to drive a sine wave at those quoted powers into a dummy load would cause an over-temperature shutdown. But then, I know you know that

Not sure where 660W comes from

It comes from the linked Crown Excell data I gave a link to. And the heatsinks are force air cooled. My C-Audio GB602 (The Crown version is the XLS602) cheerfully produces 600W per channel into 4 ohms with the fan very much keeping control (I replaced the fan when I bought it). And I'm using it at full power sine wave into 4 ohms at 20kHz in a non-audio application.

Sound reinforcement professional amps are intended for use in bars, discos and band use, with an estimated duty cycle of 0.4 to 0.5 for hours ever day. These are not hifi products - they are super rugged workhorses.


Craig

[Herald1360]

It would be interesting to see the magnetising current waveform. If it's starting to get "pips" on the peaks then saturation is near and inrush current peaks could hit 100x flc for the first few cycles at switch on.

[McMurdo]

Like I say, I've never had the bench mcb go out, ever. It's a hager type B. I regularly power-up amplifiers far bigger than this one, plus the welded mains switch tells me the problem has occurred elsewhere.
The rear mounted breaker on the amp is a 12A thermal one only.
We do have a good low impedance supply, and have our own 11kV transformer in the yard. Racks of these or similar amps are often powered from a single 32A radial switched by the fire alarm contactor, so one amp tripping the lot is not a great feature of an installation.
As an experiment I've fitted another NTC, 10ohms, directly in line with the tx primary and it allows direct powering of the amp, the bench ammeter showing 2 near-identical current pulses (FSD on the moving iron 8A meter!) for the switch on and bypass phases.

[Radio Wrangler]

Two near identical pulses it the best you can do with this arrangement, without going to a multistage soft-start.

I hope it's enough.

David

[joebog1]

I agree with mhennessy!! Check the idling current without any load.
Toroids DO take huge currents at switchon. About 10 times the normal current when its running in fact. ( worst case )
An EI transformer inrush is about 4 times running current. These figures are for "normal" transformers, that we play with, not distribution transformers.

Joe

[Glowing Bits!]

Just had a thought.
How come Kevin is having so much trouble with a small amp on a 32A supply, yet approx 10 amplifiers in flight cases can happily operate on a single phase 32A supply?
Total output for said amplifiers would be at least 10KW or more, including stage monitors.
Before anyone asks, it is absolutely shattering volume levels, not to mention the air movement in front of the speakers themselves!

[broadgage]

Originally Posted by Glowing Bits!

Just had a thought.
How come Kevin is having so much trouble with a small amp on a 32A supply, yet approx 10 amplifiers in flight cases can happily operate on a single phase 32A supply?
Total output for said amplifiers would be at least 10 Kw or more, including stage monitors.
Before anyone asks, it is absolutely shattering volume levels, not to mention the air movement in front of the speakers themselves!

IME, racks of amplifiers often have either a sequential start system, or are turned on by hand one at a time.
A true, RMS output of 10 Kw implies an input of about 20 Kw, well in excess of 32 amps single phase.
The quoted 10 Kw output might be either a very short term peak, or might simply be an advertising claim rather than actual measured watts.
(think about the "800 watt" car audio systems that work from a 30 amp fuse at 12/14 volts)

[mhennessy]

Originally Posted by Craig Sawyers

It comes from the linked Crown Excell data I gave a link to.

I disregarded that spreadsheet because it doesn't makes sense:

Starting with the current, I assume you're referring to cell C48 and/or J48?

Both of those are of the order of 22A. But looking at row 40, the first figures are for 120V/60Hz working, and the second claims to be for 220V/50Hz. That can't be right!

Moving right, we have 2 more groups of figures.

So in Q48, we have 13.6A. This apparently is 220V/50Hz as well.

Then in X48, we have 12.5A. But this is still 220V/50Hz

I was confused by this earlier, and remain confused now. Obviously these 4 groups of figures are incorrectly labelled - it is illogical to have 4 groups, the first of which is 120/60, and the next three are all 220/50 - yet all those 3 groups contain different numbers!

I don't know what they actually intended to write, but perhaps 110/60, 120/60, 220/50 and 240/50? That's just a guess...

As I say, I disregarded it and did my own "back of the envelope" calculations for full power, which of course was an extreme to make the point as it will never be expected to deliver full sine wave power - the mains transformer sees to that; just look at the sag between 8 and 4 ohms. What do you suppose its VA rating is?

However, if we believe the spreadsheet, and think we have understood the error with the group titles, then we can see that the current draw on this side of the pond is around 12-13A, not 22.5A. Still high for that transformer, which is clearly not 3kVA - this can only be a short-term rating.

(Out of interest, I looked up a 3kVA toroidal transformer (RS 752-9371), and it weighs 21kg. The whole amplifier weighs 19.5kg! It's also 125mm high, which might be a bit of a squeeze in a 2U case.)

This is based on 1/3 power operation, but that almost as unrealistic as full power. As they say themselves, it's "typical of program (sic) material at extreme clip". Music rarely has a PMR of only 10dB, and certainly not live music. They suggest that 1/8 power is more typical, and that's nearly 20dB, which is heading in the right direction...

But that aside, it still doesn't make complete sense. The 660W you cite is a third of 2kW into 4 ohms, both channels driven. But as I noted earlier, this amp can't do that - the 4 ohm power is only 1350W. So really, the 4 ohms result is a half-power test, not third-power...

Originally Posted by Craig Sawyers

My C-Audio GB602 (The Crown version is the XLS602) cheerfully produces 600W per channel into 4 ohms with the fan very much keeping control (I replaced the fan when I bought it). And I'm using it at full power sine wave into 4 ohms at 20kHz in a non-audio application.

I've no experience with your specific amp, but all PA amps that I've had on the bench will barely break out into a sweat with music that is just into clipping, but will eventually go into thermal shutdown with sine waves sharing the same peak voltage. And that's not surprising, because no PA amplifier should ever be expected to deliver their full sine wave power continuously because they would be seriously over-engineered for their intended application - not welcome in an area where weight is the primary specification! Followed by cost, sadly. Of course you have to build in some "thermal headroom" to account for high ambient temperatures, blocked filters, clogged heat sink fins, knackered fan bearings, etc, and that headroom results in a reasonable delay before the thermal cut-out works. But you do have to keep an eye (well, a thermocouple!) on the mains transformer during testing... Those are rarely monitored separately, but that's fair enough because - sorry to repeat myself - you'd never expect a PA amp to deliver a full power sine wave for any length of time. Any idea what the VA rating of the transformer in your amp is?


(As I'm sure you'll appreciate, I'm simplifying a bit because the waste heat vs % of output power is not straightforward, and full power wastes less heat than 1/2 power for a typical class B amp. For class G it's harder to predict as it depends on the number of levels and the voltages chosen.)

Originally Posted by Craig Sawyers

Sound reinforcement professional amps are intended for use in bars, discos and band use, with an estimated duty cycle of 0.4 to 0.5 for hours ever day. These are not hifi products - they are super rugged workhorses.

I do have a passing familiarity with the species, but in the present situation, this is becoming an increasingly distant memory, sadly.

They're still built to a price though.

[Diabolical Artificer]

If the amp came fitted with a 5A fuse and the amp worked before and is now blowing it, it reasons there is a fault after the toroid; does the MCB trip with the secondary isolated from the rest of amp? What is sucking in all the current as the relay engages? My thinking the toroid is a red herring. Could be wrong and dare say you'd thought of that.

Andy.

[Glowing Bits!]

Has anyone thought of checking the tranny with a Megger?
A basic check I do when fault-finding is to use a DMM, note primary/secondary readings, this is followed by using the Megger across (usually the one that causes problems) primary, crank it hard (500v) then re-check the reading with DMM, it should stay the same if it's working properly.
If all that checks out fine, do a diode test on the rectifier, these do fail eventually, it puts a short across the secondary leading to severe overheat, a fuse should have activated by then.

[Craig Sawyers]

Looking at the schematic for the GB602, there is a temperature sensor (LM35DZ) on the heatsink. That is used to control the fan speed, and if the heatsink temperature goes above a limit, the output relays fire to disconnect the speaker and an overload LED lights. The output relays also fire after a switch-on delay to prevent thuds on switch on/off.

The soft start is two PTC thermistors (YM120D370N100) in parallel, in series with a 47 ohm (23W) wirewound resistor. That is shorted by a relay after the delay time. That is in the neutral line, presumably to keep the contacts near the coil potential.

I suspect that the thermistors are there to protect the resistor from burning out if the soft start relay fails to fire.

Craig