IRS27951 / IRS27952

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi Nigel,
to get a 2chamber bobbin I use a 'Kabelbinder' (cable tie?).
I'm just copying an chinese LLC converter, where they did it like that. Get a 4mm gap between pri and sec. Unfortunately the design uses a l6599, which is not esy to get for hobbyist (correct my spelling ;-) )
Greetings from Germany.

Thanks for the tip.
I have loads of cable ties lying around.
 
L6599

Thanks Nigel, for that link. They offer it for 3.25 pounds. The german RS offers it for 6,99€.[price checked], that's rediculous. At Mouser I bought the newer, improved A-Version for 1,55 €, but 20€ transport. Only bought five, killed one. Should have orderd more, but at that time I didn't know wether the circuit would work. Since yesterday evening the regulation, burst mode, etc. is working.This evening I'll draw power.
 
Thanks Nigel, for that link. They offer it for 3.25 pounds. The german RS offers it for 6,99€.[price checked], that's rediculous. At Mouser I bought the newer, improved A-Version for 1,55 €, but 20€ transport. Only bought five, killed one. Should have orderd more, but at that time I didn't know wether the circuit would work. Since yesterday evening the regulation, burst mode, etc. is working.This evening I'll draw power.

Try Fairchld ones. They are not pin-to-pin compatible, but have about the same specs (bootstrapped low and high side FET driver, soft start, and burst mode features). They also lot cheaper (at least at Farnell).
Also look at TME, they have the IRS27951 very cheap (half of the Farnell price).

The burst mode is only needed if you have totally no laod at idle. If you power an amp with this supply you will have some watts load, so you won't need burst mode.
Also at burst mode the ZVS operation fails (hard-switching occurs) for the first dozen cycles. In my experiments with UCC25600 due to the loss of ZVS at burst mode, the primary FETs dissipated more, than without burst mode!
(the burst mode only lowered the total idle power consumption by 30%)
 
I built up another pcb.
Fried the first 27952 but the second worked ok.

I built up another transformer a lot neater than the previous one but am getting some odd measurements, I get 43uH and 600uH. This gives a k out of range.
The pcb powers up goes to 45 volts then immediately cuts out again.
I adjusted the control components for the new Fr1 but it made no difference.

Here is a picture of the transformer the two windings to the right are laid next to each other and are the secondaries. The left hand winding under the tape is the primary. The winding on top of the tape is the aux w3inding which gave out too much voltage and blew up another chip, so I just ran the next chip dropped off 340 VDC and it worked OK.
An externally hosted image should be here but it was not working when we last tested it.
 
nigelwright7557;3353489 Here is a picture of the transformer the two windings to the right are laid next to each other and are the secondaries. [/QUOTE said:
1. Put the splitter in the middle of the bobbin. If you make more width to the secondaries, the leakage will decrease. If you put in the middel, the leakage inductance will increase.
2. The two secondaries must be wind together. Pick up two wires, solder each's start to the pin, then wind the two together. (otherwise the two secondaries will have different leakge)
3. Dont use thick wire. Use wire diamater <0,5mm and trand (twist) together more wires. If you would need 1mm for that current twist together 4x0,5mm wire.
 
I rewound the transformer with the secondaries on top of each otherQUOTE]

Not top of each other, not next to each other, but together.
The better word is bifillary. This cause the best coupling between the secondaries.

I did as you suggested.
33 turns primary, 9 turns secondaries wound bifilar.
I get 415uH and 53uH.

Reading through the app note I should be getting 115uH leakage not 53uH.

I am only getting a couple of amps out before the transformer starts vibrating and the regulation dropping out.
 
I don't suppose the using of longish leads and crocodile clips could be causing problems? I can't imagine that this would be a great setup for higher frequencies and high currents. Maybe this could be causing stability issues or problems with the voltage regulation/feedback. Of course you've shown that the values expected for the transformer aren't quite right and it's most likely this, but have you tried it with shorter leads?
 
I don't suppose the using of longish leads and crocodile clips could be causing problems? I can't imagine that this would be a great setup for higher frequencies and high currents. Maybe this could be causing stability issues or problems with the voltage regulation/feedback. Of course you've shown that the values expected for the transformer aren't quite right and it's most likely this, but have you tried it with shorter leads?

The crocodile clips are just used to testing leakage/main inductance not when the circuit is live.

I had an ETD49 transformer so I put the same windings on that and get the same results as the etd54. I am only getting about 53uH leakage when I need at least a 100uH.
Gapping the transformer makes no difference to leakage inductance.
 
The crocodile clips are just used to testing leakage/main inductance not when the circuit is live.

Ah, that makes more sense.

I am enjoying this thread greatly though, learning to build a simple, but regulated SMPS is always something that I've wanted to do, but have never attempted. This type of supply appears to do and be just that and this thread has turned out to be a great resource towards anyone who is interested.

The bit that has always put me off is the part where you make your own transformer and for good reason too it seems, they are critical towards the success of the design and lori's input in this regard has, at least to me, been extremely helpful in helping to bring some clarity to the problem.
 
Ah, that makes more sense.

The bit that has always put me off is the part where you make your own transformer and for good reason too it seems, they are critical towards the success of the design and lori's input in this regard has, at least to me, been extremely helpful in helping to bring some clarity to the problem.

I have made about 6 transformers now.
The first I had the secondaries wound on top of the primary which is wrong if you want decent leakage inductance. A split bobbin is the way to go.

I was completely thrown by a cheap LC meter I have giving nonsense readings. It told me I had 10uH leakage when I had 100uH.
I found the best way to measure it was to apply through a resistor a sweeped waveform looking for the low point. I then measured the frequency on the scope and could work out the inductance from that.
L=1/ 4 PI PI F F C

I was advised to wind the 2 secondaries bifilar to keep the inductances the same.

The maths is quite hard going in the app notes.
lorylaci gave me a spreadsheet to make it much easier.

For some reason I have had terrible troubles with the 27952 SMD device. I have blown up a few now. I am not sure if the yare being cooked or it is a static problem.
I ma finding only 1 in 3 work which is a nuisance.
I bought a smd heat gun which I found just overcooked the 27952's but it has come in handy for removing them from the pcb.
I try to solder the legs as quickly as I can to reduce thermal shock.
 
Last edited:
I have got a bit further tonight.

I found reducing Rmin stopped the transformer buzzing.
But I am still not getting anywhere near 10 amps, I get 3 and thats with the output voltage sagging.

How much is the pirmary votlage when you load the secondary?
If your primary sinks to much, then regalutaion will fail.

Also you must keep in mind, that reducing the Rmin will increase your minimum frequency, and so regaulation will also fail.

The buzzing of the transformer is simply caused by the forces between the wires (F=IxB). If you wind the wires tighther, it will buzz less.
 
How much is the pirmary votlage when you load the secondary?
If your primary sinks to much, then regalutaion will fail.

Also you must keep in mind, that reducing the Rmin will increase your minimum frequency, and so regaulation will also fail.

The buzzing of the transformer is simply caused by the forces between the wires (F=IxB). If you wind the wires tighther, it will buzz less.

I am getting 320 volts on the primary side.
I added another capacitor but it made no difference. I have 440uf total.

If Rmin gets too large then the transformer starts to buzz and then the over current on the 27952 kicks in. I guess it is going into ZCS mode, taking too much current and tripping out.

I have juggled Rmin the best I can and I get 5amps max now.

I shorted out my thermistor and common mode choke in case they were dropping volts but it made no difference.
 
I checked the spreadsheet against all my components and I found my transformer had more turns than suggested by the spreadsheet.

I adjusted the transformer and now I am getting 6 amps per secondary.

The transformer no longer buzzes.

I adjusted the control components for the new leakage inductance.

It powers up mostly now but still sometimes the 27952 resets on power up.

I just need to decrease Rmin a little to stop the resetting.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.