| sibyseb |
Hi everybody, I am relatively new to this forum and something of this kind was what I have been looking for to sort out my problem. Let me start with the fact that I am not a professional or expert in electronics, but has got great interest in it and I used to do many hobby projects. The best test equipment I have is just a digital multimeter. :)
I recently bought two SMPS lead battery chargers (through ebay!) with 15A rating. They both were getting hot even without any load and with load of few amps, they became extremely hot to touch and one of them just blew up when left for 30 minutes on 4A. I found out that both of its mosfets, mains bridge rectifier and mains fulse had been blown out. :hot: :confused:
But that led me to analyse the circuit. In theory it looks fantastic and robust but practically other way round. I think they are current limited, so even shortcircuiting the output should be harmless! I am attaching the ouput section of the circuit (my apology for the low quality drawing) which took me 2 hours to figure out. The module's aluminium body (19cm x 13 cm) is attached to heat sink, which should be able to disspate several watts easily. Still the whole thing beomes very hot to touch and the output voltage becomes inconsistent.
Could anybody figure out the potential flaw with this module? Could it be due to the transformer based switching of the mosfets, which causes them to be in the ohmic phase during part of the cycle disipating heat? :confused:
I shall try to post a picture of the module as well
Thanks. |
|
|
| sibyseb |
| In continuation with my previous post, I am attaching a picture of the smps module. |
|
|
| Tolik |
Hello,| quote: | | I think they are current limited, so even shortcircuiting the output should be harmless! |
Yes, the OUTPUT is current limited, but what about primary`s current ? The schematic attached looks like pure voltage mode control, that unable to place pulse by pulse current limit of primary`s / MOSFET`s. Its mean that you unprotected against trafo`s core saturation...
| quote: | | Could it be due to the transformer based switching of the mosfets, which causes them to be in the ohmic phase during part of the cycle disipating heat? |
Absolutely NO, in case that gate driving trafo designed well. I saw 250KHz SMPS that works just fine with such setup.
| quote: | I am not a professional or expert in electronics, but has got great interest in it and I used to do many hobby projects. The best test equipment I have is just a digital multimeter...
But that led me to analyse the circuit.
|
Switching based power electronics is subject to years of learning, the experience of building hobby projects is very appreciated, but it`s not enough here. As well a digital multimeter can`t help a lot.
If you wont to deal with SMPS - 20MHz (at least) SCOPE IS MUST TO HAVE ! You MUST TO SEE YOUR WAVEFORMS to understand what`s going on in your circuit. I`m self learning SMPS from 2002, I know several engineers that designing SMPS for our and American army needs, but I didn`t saw even single man that was able to deal with SMPS without adequate measuring equipment.
So about your SMPS - there is several possible reasons for fault :
1. your trafo is saturating.
2, dead time is too sort, less than trr of rectifiers.
3, incorrect control of MOSFET`s gates ( I`m not sure about it, but still be possible )
To be sure and help you , I (and other people as well ) have to see waveforms of voltages and currents...
Good scope ( don't mean too expensive ) will be your best helper in many other projects...
Ps. output from RS seems to be misswiered... |
|
|
| sibyseb |
Thanks for your reply Tolik. All your comments are much appreciated.| quote: | | you unprotected against trafo`s core saturation... |
True. But if the trafo core is getting saturated even on minimal load condition, the whole circuit may need redesign, which is not an easy task for me. :confused: | quote: | | Switching based power electronics is subject to years of learning |
Well said, I am learning your point now. :)
| quote: | | If you wont to deal with SMPS - 20MHz (at least) SCOPE IS MUST TO HAVE |
Hmm.. much more expensive than the SMPS...:hot: Probably SMPS is not something for my present work up.
Anyway, I was thinking if there was any gross fault with the circuit which had a simple remedy.... Never mind ;) |
|
|
| Tolik |
Hi,| quote: | | But if the trafo core is getting saturated even on minimal load condition, the whole circuit may need redesign |
Saturation is not always load related.
.................. 10^8*V*Ton
B Gauss = -----------------
................... n*Ae cm^2
From the formula above can clearly be seen to what B (flux destiny) mostly related . n - number of turns, V - voltage applied & Ae - core cross section are given. The Ton, length of pulse is product of F switching & duty cycle. By playing around Fsw or duty cycle length you will be able to reduce B in core.
| quote: | | Hmm.. much more expensive than the SMPS... Probably SMPS is not something for my present work up. |
Scope good not for SMPS only, but for ulmost all electronics related tasks. So $150...200 is resonable price of abilyty to learn & to create you own designs.
| quote: | | the whole circuit may need redesign, which is not an easy task for me. |
But this is very good way to learn. Take your time...
If you still be intresting in SMPS - download & read:
http://lord-n.narod.ru/download/boo...ronics/SPSD.rar this is SMPS bible !
To open files you will be need WinDJvu program.
Hard copy of this book will cost around $120. |
|
|
| guitar_joe |
that scope has an analog bandwidth of 200kHz?!?!?!?
that would be useless for smps? |
|
|
| sibyseb |
Thanks again for your support Tolik.
From what you said, I was just wondering if this smps was designed for 110V, though it mentioned 110V / 230V autoswitching.
I am now trying how to get a scope. I shall post the results later.
I shall also read up some stuff you mentioned to 'enlighten' me.
:D |
|
|
| luka |
Hi
For both voltages! Only one jumper is nedded/not for 110/230v operation |
|
|
| Tolik |
Hi sibyseb,
I`m also think that simple SMPS as above are have not any ability to automatic selecting right mode for US/EU mains voltage... So, as Luka mentioned you may to search for appropriate voltage selector at your boards - either switch or jumper. I`m also bought 400W 13.8V SMPS from EBay, and it contain that mains voltage selector switch.
| quote: | | that scope has an analog bandwidth of 200kHz?!?!?!?that would be useless for smps? |
guitar_joe,
Let`s look at advantages & disadvantages of INEXPENSIVE scope.
1. F cut-off of SMPS feedback loop should not to exceed 5-10% of Fsw. So such scope (with additional spectrum analyzer ! ) will be good for design / analysis feedback loops for SMPS rated up to 1-1.5MHz :eek: Any DMM really useless here, even my $3000 scope not contain analyzer.
2. To power most of diy projects, even 1-2KW car amps, 25-40KHz SMPS enough. If you wont to push your trafo to its maximum, to get the best power/volume rating possible - as seen on military or aerospace designs - go for high speed digital storage scope. Personally I prefer to add $3-10 for bigger toroid, than to replace $3K scope with $5K scope + $20K trans-impedance analyzer, for getting 100-200W extra from my designs.
3.Transformer`s / choke`s core saturation is extremely FATAL for most SMPS topologies. It`s a law. Is anybody here know MORE ACCURATE & LESS EXPENCIVE device than scope to diagnose a core saturation? You probably will tell - "OK, but kit discussed DON`T contain current probes, & all other inexpensive scopes too as well. The new current probe of good brand like Tektronix or Alginet will cost 2-3 times more than kit above !"
Yes, you will be right. But as I mentioned above, there is a way to save several hundreds of $$$. Small toroid + 1..3A Schottky diode + 2(3) resistors+ some wire (+1 small capacitor)+some of you patience & work & basic soldering skills= CURRENT PROBE FOR ANY SCOPE for less than $20.
Disadvantages available as well:
1. 200 KHz not enough to analyse parasitic HF oscillations - this scope not for very high power, high Fsw designs. Really bad that will be difficult to design snubbers.
2. Not so user friendly - all additional values of signal as Ton, RMS, Duty cycle, Burst width, etc you will be need to calculate yourself. Very limited input voltages range (20V for the scope above) Additional voltage dividers are required.
3. Last, but very important too - probes quality. Actually I mean capacitance here. I don't sure that probes of this scope are in 10...30pF class.
All pro`s & con`s above, are coming for price similar to price of spare set of two 200MHz passive probes, or spare battery for scope like mine or Luca`s ;) The scope above, is still be good example, how even inexpensive scope helpful can be. |
|
|
| impuls |
| Hi, first read value of the resistor on pin 6 on TL494 and value of the capacitor on pin 5 of the Tl494. Maby is your frequency is to small. |
|
|
| sibyseb |
| quote: | | as Luka mentioned you may to search for appropriate voltage selector at your boards |
I could not locate any jumper or switch on the board for the voltage selector. Are they normally incorporated in the PWM circuitry?
| quote: | | first read value of the resistor on pin 6 on TL494 and value of the capacitor on pin 5 of the Tl494 |
Interestingly I found the resistor was 1K and capacitor 2200 pf (labelled 222 on it). I calculated the frequency of the oscillator to be 455 kHz (correct me if I am wrong). I thought that was too high for that circuit and decided to fiddle with the resistor value empirically. I found that increasing the resistor value made the mosfets less hot. With 2K8 resistor, the mosfets remained cool at no load. Finally I tested with 5K resistor. The results are
1. The mosfets are cool at no load and just hot on touch with load. (Much better than previously)
2. The output voltage remains constant when no load (I set up 13.8V)
3. With load, the ouput voltage drops, but remains stable at that. For eg. at 15A load, the voltage drops to 11V but remains at that without fluctuation. I think that is pretty useful for a battery charger as it avoids very high currents when charging a fully discharged battery.
With that setting the circuit seems to be pretty usable, except that I can occasionally hear a grrrrrr... sound from somewhere on the board when I load it with high output current (15A). That sound is absent when I use a 2K8 resistor for the TL494, but the mosfets are slightly more hot with that. So I left it with 5K.
I am not entirely sure what I did was wise. But things seem to have improved. Please post any comments. :) |
|
|
| luka |
| quote: | | I could not locate any jumper or switch on the board for the voltage selector. Are they normally incorporated in the PWM circuitry? |
They are between input rectifier and main caps
 |
|
|
| Tolik |
| quote: | | I am not entirely sure what I did was wise. But things seem to have improved. Please post any comments. |
Why not ?
Lets see TL494 datasheet from Motorola.
At page 2, clearly shown a range for Timing Resistor - 1.8...500K, so resistor that you used is ok.
At pade 5, Figure 3. A relation of Fosc to Ct & Rt. For 1K Rt, & 0.0022 uF Ct, Fosc = around 450KHz.
Now just imagine looses in MOSFETs, & trafo`s core at 450/2 = 225KHz :eek:
| quote: | | except that I can occasionally hear a grrrrrr... sound from somewhere on the board when I load it with high output current (15A). That sound is absent when I use a 2K8 resistor for the TL494, but the mosfets are slightly more hot with that. So I left it with 5K. |
Be careful with it. Acoustical noise is symptom of trafo saturation or feedback loop instability. |
|
|
| sibyseb |
Thanks for all your comments.| quote: | | They are between input rectifier and main caps |
There is no swich on the board in the input section for voltage toggle. It seems to be fitted as for 230V (According to luka's schematic)
Now that I have replaced the mosfets and fuse on the blown out board and put 5k for the tl494 oscillator. It is working pretty well. Can get 15A without much problem and it is silent.
I was thinking an SMPS of this sort normally runs at 50 - 100 khz. Could it be an error from manufacturer to have put a wrong resistor, say 1K instead of 10K.
Anyway it has been a good learning experience for me. :D |
|
|
| luka |
Hi
It could be made for 230v only you never know....
I think that could happen too, it is probably from china :D
Working at 50k should be more then enough I think... |
|
|
| N-Channel |
| quote: | Originally posted by sibyseb
I could not locate any jumper or switch on the board for the voltage selector. Are they normally incorporated in the PWM circuitry?
....Interestingly I found the resistor was 1K and capacitor 2200 pf (labelled 222 on it). I calculated the frequency of the oscillator to be 455 kHz (correct me if I am wrong). I thought that was too high for that circuit and decided to fiddle with the resistor value empirically. I found that increasing the resistor value made the mosfets less hot. With 2K8 resistor, the mosfets remained cool at no load. Finally I tested with 5K resistor. The results are
1. The mosfets are cool at no load and just hot on touch with load. (Much better than previously).....I am not entirely sure what I did was wise. But things seem to have improved. Please post any comments. :) |
In almost every circuit controlled by the TL494, R(t) is usually 15kW or 18kW, and C(t) is usually 1000pF, giving an F(osc) of ~66-72kHz, or F(sw) of 33-36kHz. Car Amps, and pc SMPSs all use the same TL494 controller 'cuz it's just about the cheapest full-featured (or nearly full-featured) PWM chip on the market. Glad to see it's working OK.
Cheerz,
Steve |
|
|
|
