cool half bridge concept but:
just about the use of IR2110.
this kind of circuit is bad, it always burn when some variations occurs or mosfet burns.
to drive high and low mosfets, i've tested the silicium and the transformer.
the transformer wins many more.
it don't burn,
it don't make burn mosfet
it don't burn the command circuit.
it can drive up to as many isolated mos you want.
at work, i have qualified a simple transformer, based on RM8 ferrite socket to drive the 4 MOS of a full bridge config. it works also with simple toroidal ferrite.
the transformer didn't burn. i burnt ~ 40 mosfet drivers before to say it is not viable.
only 10 turns are needed at 100kHz.
primary and secondaries are all the same.
just, as precautions, limit the current in primary with a 10ohms resistor, and charge the secondaries with a little charge like 10kohms.
put protection diodes at the output of circuit just before the primary.
anode on output, cathode on +Vcc
k on output, a on gnd.
for the two wires of the transformer if piloted with two output, or only on warm wire if piloted by a single output as you have.
a link is better, but some precisions again, in the figure 5, all irfd1z0 are not usefull.
http://www.irf.com/technical-info/appnotes/an-950.pdf
just about the use of IR2110.
this kind of circuit is bad, it always burn when some variations occurs or mosfet burns.
to drive high and low mosfets, i've tested the silicium and the transformer.
the transformer wins many more.
it don't burn,
it don't make burn mosfet
it don't burn the command circuit.
it can drive up to as many isolated mos you want.
at work, i have qualified a simple transformer, based on RM8 ferrite socket to drive the 4 MOS of a full bridge config. it works also with simple toroidal ferrite.
the transformer didn't burn. i burnt ~ 40 mosfet drivers before to say it is not viable.
only 10 turns are needed at 100kHz.
primary and secondaries are all the same.
just, as precautions, limit the current in primary with a 10ohms resistor, and charge the secondaries with a little charge like 10kohms.
put protection diodes at the output of circuit just before the primary.
anode on output, cathode on +Vcc
k on output, a on gnd.
for the two wires of the transformer if piloted with two output, or only on warm wire if piloted by a single output as you have.
a link is better, but some precisions again, in the figure 5, all irfd1z0 are not usefull.
http://www.irf.com/technical-info/appnotes/an-950.pdf
Hi edfed,
sorry but I don't agree with you.
If a chip burns or make other parts burn there must be a reason.
If International Rectifier sell the IR2110 chip and specifies some operating conditions it is not possible that it burns when used properly.
The IR2110 is a very popular device annd it is also widely used in the industry. For example look at the QSC power amplifiers schematics; they use the IR2110 to drive the two half brigde IGBT.
Maybe you are constantly buning the 2110, the mosfets and the PWM IC because you have huge voltage or current spikes due to layout issues.
Transformer gate drive is a nice approach but I prefer to use semiconductors instead of magnetics where possible.
@ N-channel.
Thank you very much for the IC part numbers!! I will look at the datasheets
ciao
sorry but I don't agree with you.
If a chip burns or make other parts burn there must be a reason.
If International Rectifier sell the IR2110 chip and specifies some operating conditions it is not possible that it burns when used properly.
The IR2110 is a very popular device annd it is also widely used in the industry. For example look at the QSC power amplifiers schematics; they use the IR2110 to drive the two half brigde IGBT.
Maybe you are constantly buning the 2110, the mosfets and the PWM IC because you have huge voltage or current spikes due to layout issues.
Transformer gate drive is a nice approach but I prefer to use semiconductors instead of magnetics where possible.
@ N-channel.
Thank you very much for the IC part numbers!! I will look at the datasheets
ciao
Mag-
My pleasure. Here's a more complete list of all the PWM chips I have ever worked with or know about.
Voltage Mode:
SG3525/3527*
SG3524
SG3526
TL494CN
TL594CN
TL598CN
MC34025/33025
Current Mode:
SG3525/3527*
UC1846/2846/3846
MC34025/33025
UCC3806 (BiCMOS version of '1846)
UCC3808
UC1856/2856/3856
UC1825/2825/3825**
* Can be either (See previous post), 3527 active low outputs
** Unitrode version of MC33025 - pin-for-pin compatabile
If I can think of any more 2-Ch PWMs, I will list them.
My pleasure. Here's a more complete list of all the PWM chips I have ever worked with or know about.
Voltage Mode:
SG3525/3527*
SG3524
SG3526
TL494CN
TL594CN
TL598CN
MC34025/33025
Current Mode:
SG3525/3527*
UC1846/2846/3846
MC34025/33025
UCC3806 (BiCMOS version of '1846)
UCC3808
UC1856/2856/3856
UC1825/2825/3825**
* Can be either (See previous post), 3527 active low outputs
** Unitrode version of MC33025 - pin-for-pin compatabile
If I can think of any more 2-Ch PWMs, I will list them.
edfed said:
Hi edfed,
At first I have also burned many IR2110/2113. But now I get them working properly. There is nothing wrong about IR2110/2113 ICs. If something IS burning or not working properly, then it is because of the improper use of those ICs and the fault is probably lying with the user, not IR Company.
With proper application, ir2110 works wonderfully, as I am using those from 50hz to 150kHz without any problem.
Thanks.
Two transistor forward is nice 
What is the frequency?
Two switch forward doesn't have that big core usage disadvantage compared to half-bridge when the frequency is increased into the core-loss-limited range rather than saturation limited. (Core loss is determined by flux swing p-p, not peak flux)
Also, current mode control can't be easily used on a half bridge requiring ugly hacks like balancing windings and current transformer hacks to make it stable in some conditions.
The generation of Vdrive should have it's winding turned around and D24 and L4 removed to use the flyback from the inductor instead. (I'm assuming it's low power). The way it's now makes it's output voltage very dependent on input voltage.
What is the frequency?
Two switch forward doesn't have that big core usage disadvantage compared to half-bridge when the frequency is increased into the core-loss-limited range rather than saturation limited. (Core loss is determined by flux swing p-p, not peak flux)
Also, current mode control can't be easily used on a half bridge requiring ugly hacks like balancing windings and current transformer hacks to make it stable in some conditions.
The generation of Vdrive should have it's winding turned around and D24 and L4 removed to use the flyback from the inductor instead. (I'm assuming it's low power). The way it's now makes it's output voltage very dependent on input voltage.
sure?
i don't believe you.
i burnt at least all i had, about 40 from 3 different drivers.
IR 2110
L6384
L6385
the best is with transformer to drive the mosfets, it burns very less.
but burn.
now, i prefer to work on limited voltage/current.
i have problems now to drive the transformer, but with SG2525 it is very good.
for obsolescence and cost reasons, i must do it with pics.
what the problem now?
the pic don't have enough power to drive correctlly the tranformer & mosfets.
i need at least a 5V dual buffer that can output 2 * 1A
to obtain 20 volts on the four secondies.
a pic cost 0,60 €
a SG 2525 cost 5€
what will be the price of the buffer?
tested with 2n2222 & 2n2907 in push pull, not enough power or speed.
i don't believe you.
i burnt at least all i had, about 40 from 3 different drivers.
IR 2110
L6384
L6385
the best is with transformer to drive the mosfets, it burns very less.
but burn.
now, i prefer to work on limited voltage/current.
i have problems now to drive the transformer, but with SG2525 it is very good.
for obsolescence and cost reasons, i must do it with pics.
what the problem now?
the pic don't have enough power to drive correctlly the tranformer & mosfets.
i need at least a 5V dual buffer that can output 2 * 1A
to obtain 20 volts on the four secondies.
a pic cost 0,60 €
a SG 2525 cost 5€
what will be the price of the buffer?
tested with 2n2222 & 2n2907 in push pull, not enough power or speed.
edfed said:
Hi edfed,
I still prefer IR2110 to gate drive transformer.
But if you want to use gate transformer from pic microcontroller, and your frequency is <20kHz, you can drive TLP250 directly from microcontroller and then the transformer can be driven through the TLP250 with currents upto 1.5A.
For frequencies >20kHz, you could use power transistors as buffers from the microcontroller, since the transformer driving current requirement is more and I think that is the easiest solution for you at higher frequencies.
Thanks.
Hello Megajocke
the frequency is set at 150kHz.
I don't understand your point about vdrive generation.
Vdrive is a low power output (max 50mA) that will be regulated at 12V on another board.
You mean to reverse the Vdrive winding on TR2 and remove the L5 and D22? In this case I need also to half the number of turn of the Vdrive winding; am I right?
Another point on TR2: is the direction of the two mains winding correct or not? I am not sure about it.
I have now built the circuit and of course I have some problems.
Without load (only the two 8.2k resistors in parallel to the outputs, so 10mA) it continuosly try to start and the it shuts down. After some time it starts but I continously hear a scratching noise coming from TR1 or TR2 (I don't know).
Once it starts it can drive a 50W load but after few time it stops and and retry to start continously.
I was assuming that it stops because of spikes on the current sense pin but it seems not the case.
It seems like a protection of the FAN7554 is triggering but I don't know why.
Please help
thank you
the frequency is set at 150kHz.
I don't understand your point about vdrive generation.
Vdrive is a low power output (max 50mA) that will be regulated at 12V on another board.
You mean to reverse the Vdrive winding on TR2 and remove the L5 and D22? In this case I need also to half the number of turn of the Vdrive winding; am I right?
Another point on TR2: is the direction of the two mains winding correct or not? I am not sure about it.
I have now built the circuit and of course I have some problems.
Without load (only the two 8.2k resistors in parallel to the outputs, so 10mA) it continuosly try to start and the it shuts down. After some time it starts but I continously hear a scratching noise coming from TR1 or TR2 (I don't know).
Once it starts it can drive a 50W load but after few time it stops and and retry to start continously.
I was assuming that it stops because of spikes on the current sense pin but it seems not the case.
It seems like a protection of the FAN7554 is triggering but I don't know why.
Please help
thank you
maybe a cross conduction of your transistors, creating not a spike, but a permanent current, that your sense will integrate to give a higher rms value than if running normally.
maybe your diodes are too warm, and then, they change their characteristics.
i had this problem, as the temperature grows, the little spike on widings grows constantlly, up to 600 volts (mosfet limit) if i let it do. resolved with a heat sink.i saw this problem on a converter, i presume it can create your problem.
i saw this too on a 24Vdc to 127Vac converter, and it was mainly:
1/ the current sensor that wasn't well calibrated, due to a failling sense tranformer.
2/ a bad capacitor in the current sense filter
3/ transistor cross conduction problem due to driver problem due to base capacitor problem.
as you it started ok, worked few minutes, and then, restart constantlly.
some times it will restart only one time, and sometimes, it will restart all the day & night.
but if it starts when cold, and after some minutes it will bug, it means your components need some mintos.
but before to seek a complex cause, be sure it can be a very dumb problem like this one.
maybe your diodes are too warm, and then, they change their characteristics.
i had this problem, as the temperature grows, the little spike on widings grows constantlly, up to 600 volts (mosfet limit) if i let it do. resolved with a heat sink.i saw this problem on a converter, i presume it can create your problem.
i saw this too on a 24Vdc to 127Vac converter, and it was mainly:
1/ the current sensor that wasn't well calibrated, due to a failling sense tranformer.
2/ a bad capacitor in the current sense filter
3/ transistor cross conduction problem due to driver problem due to base capacitor problem.
as you it started ok, worked few minutes, and then, restart constantlly.
some times it will restart only one time, and sometimes, it will restart all the day & night.
but if it starts when cold, and after some minutes it will bug, it means your components need some mintos.
but before to seek a complex cause, be sure it can be a very dumb problem like this one.
What's wrong with that? He has winding to make Vaux which is then regulated and put on IR, I have the same way, just no regulation of this voltage and it is in no way current limited...I would provide more then just 50mA for IR, if you have 150 or even 200m available, then you will have no problems with charging it, if you even have any
Hi Luka,
is not a matter of powering on the 2110 is a matter of charging the bootstrap capacitor.
This capacitor has the positive plate connected to the bootstrap diode and the negative plate connected to the source of the high side switch.
In a half bridge design when the upper mosfet is on the lower mosfet is off, in a dual switch forward the two mos are on at the same time. This means that the bootstrap capacitor has not a DC path to ground when the high side mosfet is off (because the low side mosfet is off too) so it can not recharge.
Thank you areza for pointing me out that trick on the app note. I have already tried on my board but I still have problems. I think I have a sort of cross conduction between the high side mosfet an the additional mosfet I have added to recharge the bootstrap capacitor. The high side mosfet is heating up.
I have checked the waveform between the high side gate and the high side source and I find squarewaves of only around 5V amplitude. This means that the bootstrap capacitor is not recharging properly. (basically the same behaviour with and without the circuit in the app note). Probably the value of the bootstrap cap is too high (10uF), I have calculated it using the equation in the app note and I find 250nF. I will try to replace the cap with a smaller one and see whats happens.
By the way the SMPS is working now, I have some instabilities on the startup circuit (now fixed). I can continously power a 50W load but the upper mosfet is heating up (the lower stays cool).
thank you
is not a matter of powering on the 2110 is a matter of charging the bootstrap capacitor.
This capacitor has the positive plate connected to the bootstrap diode and the negative plate connected to the source of the high side switch.
In a half bridge design when the upper mosfet is on the lower mosfet is off, in a dual switch forward the two mos are on at the same time. This means that the bootstrap capacitor has not a DC path to ground when the high side mosfet is off (because the low side mosfet is off too) so it can not recharge.
Thank you areza for pointing me out that trick on the app note. I have already tried on my board but I still have problems. I think I have a sort of cross conduction between the high side mosfet an the additional mosfet I have added to recharge the bootstrap capacitor. The high side mosfet is heating up.
I have checked the waveform between the high side gate and the high side source and I find squarewaves of only around 5V amplitude. This means that the bootstrap capacitor is not recharging properly. (basically the same behaviour with and without the circuit in the app note). Probably the value of the bootstrap cap is too high (10uF), I have calculated it using the equation in the app note and I find 250nF. I will try to replace the cap with a smaller one and see whats happens.
By the way the SMPS is working now, I have some instabilities on the startup circuit (now fixed). I can continously power a 50W load but the upper mosfet is heating up (the lower stays cool).
thank you
nice its working, looks like there is few pin left in the transformer , it would be good to add a winding made with a diode resistor and zener for charging up bootcap , this way i guess even a 100uf would not be a problem and the bootcap charging would be out of the high current path and not through the lower fet and current sens resistor,
Hi Areza,
I have added 3 turns on my transformer and I use this winding to charge the bootstrap cap.
I have connected one side to Vs and the other side to the capacitor trough a diode and a 10R resistor. I have clamped the voltage on the boot cap at 18V with a zener diode.
The mosfets does not heat up anymore, even with around 100W load.
I still need to stabilize the feedback loop. I still hear nasty sound coming from the transformer. I have tried to change the value of the compensation cap on the TL431 and the sound changes; I think that this is related to the feedback stability, am I right?
thank you.
I have added 3 turns on my transformer and I use this winding to charge the bootstrap cap.
I have connected one side to Vs and the other side to the capacitor trough a diode and a 10R resistor. I have clamped the voltage on the boot cap at 18V with a zener diode.
The mosfets does not heat up anymore, even with around 100W load.
I still need to stabilize the feedback loop. I still hear nasty sound coming from the transformer. I have tried to change the value of the compensation cap on the TL431 and the sound changes; I think that this is related to the feedback stability, am I right?
thank you.
Yes, you are right.
You only need now to test with a slow lowpass filter.
If the feedback is too fast, on the scope, you'll see something like a chaos due to instable command.
At each commutation, the rectifier diodes on secondary create pulses.
These pulses will oscillate, depend on speed of your diodes, very fast recovery diodes required there, mounted on a heat sink.
And then, the feedback will be instable. Don't hesitate to overdimension your components.
You only need now to test with a slow lowpass filter.
If the feedback is too fast, on the scope, you'll see something like a chaos due to instable command.
At each commutation, the rectifier diodes on secondary create pulses.
These pulses will oscillate, depend on speed of your diodes, very fast recovery diodes required there, mounted on a heat sink.
And then, the feedback will be instable. Don't hesitate to overdimension your components.
Thank you edfed,
I will try it and let you know.
I don't understand what you mean when you are saying "on the scope you will see something like chaos": where should I probe with the scope?
In attached you will find an updated version of the schematics with the modifications I have done so far.
Please comment on it and let me know
ciao
-marco
I will try it and let you know.
I don't understand what you mean when you are saying "on the scope you will see something like chaos": where should I probe with the scope?
In attached you will find an updated version of the schematics with the modifications I have done so far.
Please comment on it and let me know
ciao
-marco
mag said:
If you just reverse it right now you will have about 20V regulated on the aux output as long as there is some pretty small load on the main outputs. (Comes from turns ratio on the inductor) When primary switches are off TR2 will be pulling current through D17 with negative at dot-end and 85V+v(D17) over 4-1. This is divided by turns ratio and peak rectified then.
The main windings are correct with respect to each other.
Hi all,
I have stabilized the feedback loop but now I still have some probelems with the current mode control.
As soon as I increase the load the smps start to continously turn on and off.
I have measured the voltage on the Isns pin of FAN7554 and I see very high peaks even with 10mA load.
I have captured with the scope the waveform across the current limit resistor (in attached). Those spikes are huge (from +30V to -20V with no load). I suspect that I have too much inductance on the current sense resistor (0.22ohm, 3W, wirewound!!).
Should I replace the resistor with something not wirewound?
Slowing down the mosfet gate charge time (increasing the gate resistor) can be effective?
thank you
I have stabilized the feedback loop but now I still have some probelems with the current mode control.
As soon as I increase the load the smps start to continously turn on and off.
I have measured the voltage on the Isns pin of FAN7554 and I see very high peaks even with 10mA load.
I have captured with the scope the waveform across the current limit resistor (in attached). Those spikes are huge (from +30V to -20V with no load). I suspect that I have too much inductance on the current sense resistor (0.22ohm, 3W, wirewound!!).
Should I replace the resistor with something not wirewound?
Slowing down the mosfet gate charge time (increasing the gate resistor) can be effective?
thank you
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.