| RiskCord |
Hello guys.
I'm currently designing the power stage for a Class D amplifer. My primary choice at the moment are the IRFB4020pbf power MOSFET.
Concerning this device, even when internal body diode specs are relatively good, they obviously look poor when compared to discrete ultrafast soft-recovery diodes or schottky rectifiers.
I was wondering about disabling the internal diode by means of a series and an antiparallel diode. I'm aware that conduction loses would be increased and also strait inductances but it could pay off in term of relaxing MOSFET current stress and minimizing switching loses... at least at first sight. I saw a couple of designs around the web and it called my attention that the antiparallel diode is usually a ultrafast diode and not a schottky as I would have expected (because of they not presenting reverse recovery issues and having smaller Vf). Would it be because there is not too much availability of schottky rectifiers with high reverse voltage capability?
Anyway, if any of you have tried this topology, comments on what resulted would be very welcome!
Thanks!
Pablo |
|
|
| Tekko |
| All you have to do is place an even faster diode in parallel with the fet, then the internal diode wont do anything. |
|
|
| Pafi |
| quote: | | the internal diode wont do anything. |
Are you sure? Check forward voltages of a (hot) body diode and an ultrafast or schottky diode!
The voltage drop across FET must be increased. Series diode or resistor, then FRED parallel with them. |
|
|
| Dave |
| I agree with Tekko, the usual practice is a schottky diode across the FET. ie. in parallel with and pointing in the same direction as the body diode. |
|
|
| Workhorse |
I have used seires schottky and anti parallel Fred with mosfets to get rid of body diode issues which are posing reliability problems.
Losses are not that high, only slightly more, but again what you get with this approach is bullet proof reliability from the drastic effects of freewheeling current.
Have a look at this combo device
http://ixdev.ixys.com/DataSheet/16f...fc143100898.pdf |
|
|
| JohnG |
It will be hard to find a Si Schottky with a 200V rating, and it is likely that its voltage drop will be comparable to the MOSFET. You can look at SiC or GaAs Schottky diodes, but their voltage drop is likely to be even higher.
The voltage drop of the external anti-parallel diode should be substantially less than that of the internal body diode. |
|
|
| Lars Clausen |
| Workhorse: Hi nice to see you again. Do you have any estimate on the impact on the natural THD of the output stage, from your circuit, as compared to a direct connected MOSFET? |
|
|
| Workhorse |
| quote: | Originally posted by Lars Clausen
Workhorse: Hi nice to see you again. Do you have any estimate on the impact on the natural THD of the output stage, from your circuit, as compared to a direct connected MOSFET? |
Nice to see you too, Lars !!
No, I never tried that open-loop test, in my close-loop design the THD is not effected by using series schottkys, but one thing I have achieved is added reliability with this scheme. ;)
My amp runs with rails @ +/- 185VDC and target application is professional use..........so i always focus for reliability first, i donot care whether the THD is 0.1% or 0.001%, because it hardly matters in sound reinforcement systems..........but Yes, it does matter in fancy amps:cool: |
|
|
| Lars Clausen |
+-185V :dead:
No we don't go that high on fancy amps :clown: |
|
|
| RiskCord |
| Thanks Workhorse for your comments. The amplifier I am designing will work in +- 75. Then the impact of freewheeling currents will be quite smaller than in your case, but it would be quite interesting to compare both performances with and without series schottkys. Sadly I dont have the time to build both prototypes so I guess I'll have to pick one in advance. I'll run simulations as accurate as possible and decide on this basis... |
|
|
| Pabo |
Tekko and Dave
Just consider that the current has to leave the channel and enter the parallell diode in a matter of nano seconds if it is going to give an improvement on reverse recovery. If the parallell diode has 0,2V lower drop and the stray inductance and internal inductance inside the MOSFET/diode is 10nH you can easily calculate that it takes 20A of current 1 microsecond to "move" to the other diode. Has anyone here seen an improvement on EMI, efficiency or THD by doing this? |
|
|
| Pafi |
Pabo!
IMHO your post a little difficult to understand, I think you could tell it in a simpler way: the diode with smaller forward voltage drop will conduct most of the current, hence it will determine recovery speed. And this will be the body diode in this case, if you just parallel them. |
|
|
| Eva |
| Forward voltage drop is not the only parameter that matters, actually parasitic inductances also determine current sharing between the body diode and any external diode. |
|
|
| Pafi |
| You are saying something! |
|
|
| Pabo |
Pafi
The body diode only conducts after the channel has been turned off if the current is flowing from source to drain (i.e. freewheeling). Immediately when the channel is turned off the current will flow through the body diode. This is of course a simplification as the body diode is actually the channel itself when not turned on. In order for the current to move to an external diode which is connected in series a certain time is needed. In a class d amplifier it has to happen within the dead time which is usually very short. So if the voltage drop is even half of the body diode it will take much longer than the dead time itself hence giving no improvement in reverse behaviour.
I am missing something? |
|
|
| peranders |
| I have tested this with 475 A and a separate diode didn't help much so you are pretty much right. |
|
|
| Pafi |
| quote: | | I am missing something? |
Yes, the resistance of the channel. The whole reverse current cannot flow on the channel typically. The output current can be more then 10 A, the Rdson can be more then 0,12 ohm (-> 1,2 V drop), so body diode can conduct during freewheeling easily. |
|
|
| Pafi |
peranders!
What is "475 A"?
And how could you test this: | quote: | | if the voltage drop is even half of the body diode |
Where did you get from a >200 V hyperfast diode with <0,4 V forward voltage? |
|
|
| Pabo |
Pafi
OK, in that case I suppose it helps. I usually use FETs with much lower rds and if the rds is that high the body diode is usually OK. |
|
|
| phase_accurate |
Has anyone ever tried to use a very small series resistors in the drain path ? This should also help to reduce reverse-recovery stress. It does however not speed-up anything but it would help to reduce the current-peak. If dimensioned properly it would not give more conduction losses than a series diode.
Regards
Charles |
|
|
| ChocoHolic |
| quote: | Originally posted by RiskCord
Thanks Workhorse for your comments. The amplifier I am designing will work in +- 75. Then the impact of freewheeling currents will be quite smaller than in your case, but it would be quite interesting to compare both performances with and without series schottkys. Sadly I dont have the time to build both prototypes so I guess I'll have to pick one in advance. I'll run simulations as accurate as possible and decide on this basis... |
..as accurate as possible?
How about trying as accurate as necessary?
It is allowed to neglect quite a lot of things, but I would like to highlight two crucial points that are most often not simulated, but tend to show their dragon face in reality.
- The parasitic layout inductances and the parasitic inductances of the devices.
Nice start for TO-220 is to simulate roughly 5nH in every leg in order to include the unavoidable geometric properties of that package. Also put about 0.5nH inductance for every mm copper trace in your high current paths (MosFets, free wheeling diodes, rail caps....).
- Your diode models should incoporate the reverse recovery characteristic.
And then simulate hard switching conditions, means go for output currents above the filter ripple current... Force PSPICE to go for small time steps (<10ns), otherwise you might miss the clue... |
|
|
| RiskCord |
As accurate as necessary... I like that better.
Yes, I included the parasitic inductances, the non-linear Gate-Drain and Drain-Source capacitances and I made a model representing the reverse recovery characteristics of the diode. The waveforms looks a bit ugly, so I guess the simulation is working just fine :D
Thanks for the advice on the inductance of PCB traces. I didnt know how to estimate them so I just doubled the values of the MOSFETs inductances. Anyway, I'm just about to start the PCB design and I guess I'll upload my design here so you can have some fun criticizing it :D
By the way, I'm not using PSpice. Instead, NL5 is my choice. It works on piece wise linear approximation which is quite suitable for non-linear applications as it is extremely fast and doesn't have (usually) convergence issues. |
|
|
| ChocoHolic |
| quote: | Originally posted by RiskCord
The waveforms looks a bit ugly, so I guess the simulation is working just fine :D
|
...resonances in the frequency range between 10MHz-100MHz ?
You seem to be on the right track.
;) |
|
|
| fernando_g |
| quote: | Originally posted by Pabo
Tekko and Dave
Just consider that the current has to leave the channel and enter the parallell diode in a matter of nano seconds if it is going to give an improvement on reverse recovery. If the parallell diode has 0,2V lower drop and the stray inductance and internal inductance inside the MOSFET/diode is 10nH you can easily calculate that it takes 20A of current 1 microsecond to "move" to the other diode. Has anyone here seen an improvement on EMI, efficiency or THD by doing this? |
I remember attending a technical presentation by one of the major MOSFET manufacturers (IR, Fairchild? don't remember) that explained this very well. They made the point that the only way to achieve the intended results was to co-package the Fet and the schotkky.
Of course, they also had in the market Fet-schottky devices they were trying to sell...
;) ;) ;) ;) ;) |
|
|
| mias |
Hello,
I am new here in this forum and I am very interested in this topic.
An interesting alternative to the shottky diode is a
- voltage control
- of the output voltage of the freewheel FET
- during the transition between the
- freewheel phase and the
- switching phase
If you
- control the output voltage to a value, e.g. 0.6V
- the intrinsic diode of the freewheel FET
- will not become conductive.
I attached some waveforms. If anybody is interested, I will give much more explanations.
Greetings,
Matthias
P.S.: I was posting something before. Please consider this as a draft. I was pushing the Submit Reply Bottom unintentionally. |
|
|
| lumanauw |
Hi, Fernando,
| quote: | | I remember attending a technical presentation by one of the major MOSFET manufacturers (IR, Fairchild? don't remember) that explained this very well. They made the point that the only way to achieve the intended results was to co-package the Fet and the schotkky. |
What part #/manufacturer is that FET/Schottky? The IXYS one is very difficult to get. |
|
|
| lumanauw |
Hi, MIAS,
| quote: | | An interesting alternative to the shottky diode is a....... |
Very interesting. Could you tell more?
In your right drawing, there is additional C at the base of totem pole and R+dioda from base of totempole driver to drain of mosfet.
What are these for? |
|
|
| classdphile |
| It's just a snubber, common technique to slow the switching transition, long since demonstrated on this forum as well. |
|
|
| Pafi |
Not really. It's there to control on-state Vds. It works very similar to schottky transistors.
The problem is it is useful only if Iout*Rdson<0.5 V, wich is not usual in high-power ClassD amps. |
|
|
| RiskCord |
"What part #/manufacturer is that FET/Schottky? The IXYS one is very difficult to get."
Fairchild has a series of MOSFET/Schottky combos but their all for low Vds. (< 60v)
Microsemi has its MOSKEY series, they go up to 100 v (MSAER57N10A) but it's the kind of custom package that Eva would hate :D
And finally, IRF has its Fettky series (cool name, isnt it?!), but again, low voltage applications.
Obviously the reason for this is that physically it is not an easy task to develope schottky diodes with a large breakdown voltage. In fact, discrete schottky diodes with Vbd > 100 are hardly available.
On the other hand, for this range of Vds I think you can get conventional MOSFETs with very nice body diode specs. In any case, if time and money is available, it wouldn't hurt to try both and compare performances.
Regards
Pablo |
|
|
| lumanauw |
Hi, Pablo,
You're right, in low voltages the new mosfets are already good. The problem is for 300-400V mosfets. IXYS one is very good, but digikey don't stock them.
I wonder if there are other manufacturer that produce 300-400V mosfet+series schottky+parrarel ultrafast dioda in one package, and "buyable" :D
MIAS' method is interesting. I just wanted to know what happens with this technique if the current in the inductor is 10's of ampere headed to base of the totempole's predriver (only mA here). Will it ruin the predriver or stage before predriver? |
|
|
| mias |
Hello lumanauw,
| quote: | | Very interesting. Could you tell more? |
I made this circuit in my vacancy. The really funny thing is: It looks risky, but it is not. The thing is a
- MOSFET
- can get conductive in reverse mode
- below its threshold voltage.
This why you can control a MOSFET with a gate voltage very little above the threshold voltage. And this is why during the whole transistion a bridge short circuit is safely prevented. As one can see in the graph - when the MOSFET gets into forward mode the gate-source voltage is below the treshold voltage.:)
| quote: | | In your right drawing, there is additional C at the base of totem pole and R+dioda from base of totempole driver to drain of mosfet. |
This capacitor has the effect to eleminate the influence of the miller capacitor. In reverse mode this capacitor can lead to a ringing tendency. The gain of the control circuit is high, can be decreased with a resistor between the emitter of the npn transistor in the totempole circuit and GND. But this has then other problems - I tried it.
| quote: | | I just wanted to know what happens with this technique if the current in the inductor is 10's of ampere headed to base of the totempole's predriver (only mA here). Will it ruin the predriver or stage before predriver? |
I did not try so many AMPs. Then also the stray inductances may play a role. Nevertheless, the MOS-Drive circuit is safe from being damaged. The Push Pull Circuit has to be low ohmic in any case, otherwise you will get some funny steering of the Freewheel FET in linear mode also, when you use a conventional MOS Drive.
Greetings,
Matthias
--------------------------------------
Hello Pafi,
| quote: | | The problem is it is useful only if Iout*Rdson<0.5 V, wich is not usual in high-power ClassD amps. |
Yes, but the MOSFETs are getting better and better, the RDSOns are getting smaller and smaller...:) And also with 0.7V control there is an effect. The voltage drop of the intrinsic diode in the amp range is pretty high.
Greetings,
Matthias
--------------------------------------
Hello classdphile,
| quote: | | It's just a snubber, common technique to slow the switching transition, long since demonstrated on this forum as well. |
I am really very interested, whether this circuit is already state of the art. Do you or anybody here have any idea, where I can find this out?:bigeyes:
Greetings,
Matthias |
|
|
| Pafi |
mias!
| quote: | | The voltage drop of the intrinsic diode in the amp range is pretty high. |
Are you sure? I see less then 0.6...0.8 V on every datasheet, and even less when diode is hot.
But otherwise, how do you control low side FET when current polarity is normal (from drain to source)? |
|
|
| lumanauw |
Hi, MIAS,
I have a dumb question. In your drawing, you don't use external dioda (parrarel to the body dioda).
If the body dioda is already disabled with your method. The upper mosfet is already off, but the lower mosfet is not on yet, the left side of the inductor will instantly change polarity 180deg, where will this energi inside the inductor go to, since the body dioda is already disabled, but there is no external dioda (parrarel with body dioda's position)? |
|
|
| Iyremenko |
| quote: | Originally posted by Workhorse
I have used seires schottky and anti parallel Fred with mosfets to get rid of body diode issues which are posing reliability problems.
|
That is the ultimate way, to solve the body-diode conduction problem. :)
Other tricks are just band-aids to mitigate the problem, but it is never eliminated. |
|
|
| mias |
Hello lumanauw,
| quote: | | If the body dioda is already disabled with your method. The upper mosfet is already off, but the lower mosfet is not on yet, the left side of the inductor will instantly change polarity 180deg, where will this energi inside the inductor go to, since the body dioda is already disabled, but there is no external dioda (parrarel with body dioda's position)? |
One cannot disable the intrinsic diode. All you can try is to get this part of the freewheel FET not conductive. In the transistion phase
- when the upper FET is switching off
- and the regulator circuit is not reacting so fast,
the only thing what is happening is that the diode could become conductive, before the freewheel FET is switched
- either hardly
- or the regulation is reacting.
Hello Pafi,
| quote: | | Are you sure? I see less then 0.6...0.8 V on every datasheet, and even less when diode is hot. |
There are 3.9mOhm, D2-Pack, 100V FETs on the market. The name is IPB04CN10N G:
http://www.infineon.com/dgdl/IPP04CN10N_Rev1.2.pdf?folderId=db3a3043156fd5730115c7d50620107c&fileId=db3a3043156fd5730115c7df60f4108a
Let us asume,
- we go up to 125deg
- we have here an RDSon of 8mOhm
- and we control to 0.4V.
Then we control a current of
- 0.4V/8mOhm
- 50A at high temperatures
When we assume a supply voltage of
- 60V.
We can control a power of
- 3kW with here.
| quote: | | But otherwise, how do you control low side FET when current polarity is normal (from drain to source)? |
In the circuit diagram I sketched the MOS-driver is pretty lowohmic. It must only be active when the Low Side Transistor cannot generate any switching slopes, etc.. In case of the low side transistor gets in the role of the switching FET, then you have to steer on the FET more high-ohmic, i.e. with a defined gate resistor.
Nevertheless, I strongly recommend in any case to do this, also in case of I do not want any control. In case of you use the FET as Freewheel FET always take care you have a low ohmic - I mean a very low ohmic - current sink. Otherwise you will charge up your gate and steer the Freewheel FET in the linear region during the transistion phase, which makes neverending troubles.
Of course, in the phases where the FET gets forward conductive, you cannot switch OFF the gate without preresistor.
Greetings,
Matthias |
|
|
| Pafi |
mias!
| quote: | - 50A at high temperatures
When we assume a supply voltage of
- 60V.
We can control a power of
- 3kW with here. |
An amplifier is used to drive speakers. Speakers have a given impedance, so you can't multiple any currents and voltages to determine output power!
We prefer single-ended topology. +/-40 V allowes 200 W @ 4 ohm.
If we stay below 100 V, the body-diode is not a serious problem, there are (cheap and low Qg!) MOSFETs with very low stored charge, like IRF540Z, FDP3652, etc... But 100 V is sometimes not enough.
There are some 150 V MOSFETs with low enough Qrr, but saidly it's almost impossible to buy them here.
The real problem starts at 200 V. (Read the first post of this topic!)
| quote: | | In case of the low side transistor gets in the role of the switching FET, then you have to steer on the FET more high-ohmic, i.e. with a defined gate resistor. |
I don't understand this! High/low ohmic? How could this solve the problem? MOSFET needs completely different gate-control depending on the direction of drain-current, because at positive drain-current your control-loop turns to positive feedback, and switches off. In an amplifier this happens in every cycle at idle (and at low signal levels), but you mustn't allow switching off just because the current changes its polarity, FET have to stay ON till the next half-cycle! |
|
|
| mias |
Hello Pafi,
| quote: | | I don't understand this! High/low ohmic? How could this solve the problem? MOSFET needs completely different gate-control depending on the direction of drain-current, because at positive drain-current your control-loop turns to positive feedback, and switches off. In an amplifier this happens in every cycle at idle (and at low signal levels), but you mustn't allow switching off just because the current changes its polarity, FET have to stay ON till the next half-cycle! |
I might be wrong, because I did not try it up to now. But let me give a chance to explain.
My circuit is only active, when this signal at the entry of the totem pole stage is high because then - and only then - a current can flow through this small signal diode that closes the control lope.
When you want to steer ON the FET in forward mode, you have to put the entry signal of the totem pole drive to zero, and then this small signal diode is no longer conductive and will have no influence.
This means the circuit I propose is only active when you try to switch OFF the FET. And even then, it is only active in reverse mode. When you have a positive drain source voltage the current through the small signal diode will become zero because of the resistor in series. All the current which is coming from the input signal and the resistor between the input signal and the base of the bipolar transistor at the entry of the totem pole stage is flowing in the base of this transistor.
There is only a problem with the switching slopes. You might get a too fast di/dt with this gate drive circuit when you are operating the transistor in forward mode. Neverheless, in forward mode the control loop cannot be active. It will only operate in reverse mode.
Greetings,
Matthias
P.S.: My circuit is just for substituting the shottky diode in parallel to the FET, means for realtivly low voltages. When you have a real high voltage operation you can oly get rid of the storage charge effect by
- two diodes
- one in the same current path of the Freewheel FET
- but this one has antiseriell direction to the intrinsic diode of the FET.
- the other one parallel to the the last described diode and the the Frewwheel FET.
- the direction of this diode is same as the direction of the intrinsic diode.
Thiis solution is common.
Of course, when the RDSON*I is higher than 0.5V, you can forget about my idea. Nevertheless, MOSFETs will become better and better. And to dope silicon with gold to achive a fast recovery is not easy from the point of reliablity. |
|
|
| Pafi |
Hello, mias!
| quote: | | And even then, it is only active in reverse mode. |
OK, but my question is: what happens in forward mode? |
|
|
| mias |
Hellp Pafi,
| quote: | | OK, but my question is: what happens in forward mode? |
In forward mode the circuit behaves like a circuit with a usual MOS Drive. There is no voltage control at all.
- When you want to turn ON the MOSFET, it will be turned with maximum gate source ON voltage (e.g. 10V).
- You turn OFF the FET with minimum gate source OFF voltage (->0V)
And now a design proposal that is valid for every MOS Drive:
- with voltage control in reverse mode
- without voltage control in reverse mode.
The switching slopes could be very much too sharp. For a freewheel FET where you expect high capacitve discharge current in the amp range, it is a must to have a very low ohmic current sink in your MOS Drive Circuit. Otherwise, you will get parasitic effects. The miller capacitor could definitly charge your gate during the transistion from the freewheel to the switching phase! That can lead to driving the FET in linear mode.
When you drive the FET in forward mode, there might be the necessarity to disable this very low ohmic current sink and to discharge the gate with a defined resistor. Then you get a slower di/dt, dv/dt that leads to less noise.
Of course, when you have an antiseriell Diode in the FET path and a parall to to this Diode and the FET, you do not need to consider this miller capacitor charge effect any more. There is no need for a very low ohmic current sink.
Greetings,
Matthias |
|
|
| Pafi |
| quote: | | In forward mode the circuit behaves like a circuit with a usual MOS Drive. |
Does it? How? What input do you feed to it in order to behave like this, and what determines when do you turn it to this operation mode? |
|
|
| mias |
Hello Pafi,
I made a sketch that explains the relevant current flows in the MOS Drive circuit.
Greetings,
Matthias |
|
|
| Pafi |
Hello, mias!
Thanks, but again:| quote: | | what determines when do you turn it to this operation mode? |
|
|
|
| Eva |
| This won't work as expected with real components. The driver stage and the additional transistor are in no way fast enough. In practice crossover times may approach 20ns during high current hard switching. |
|
|
| mias |
Hello Pafi,
A) When I feed the
- input voltage
- in front of the totem pole stage
- with +5V
then the aparture effects that the MOSFET
- is OFF in forward mode
- is voltage controlled in reverse mode.
B) When I put the
- input voltage
- in front of the totem pole stage
- to GND
I will switch OFF the FET in any case.
Greetings,
Matthias
------------------------------------------------------------------------
Hello Eva,
When you suceed to get such a short transistion time (20nsec), then
- storage charge
- does not have time to bild up
- in case RDSOn * I < 0.6V.
You are trying to minimize the
- dead time between the
- freewheel and the switching phase.
When you can ensure, that you do not come into trouble and there is no risk for a bridge short, then this is fine.
My circuit has a similar effect, but the switch off comand for the freewheel FET comes a little bit earlier than the 20nsec before you give the switch ON command for the switching FET. Therefore you can call my circuit also dead time optimization circuit.
At very high switching speeds
- the stray inductances have the effect,
- that my circuit would switch OFF the freewheel FET immediatly
- when the switching FET is turning ON.
The interaction between
- the stray indances and
- the high di/dt
- will cause huge positive voltages.
In this case my circuit would not effect as a control circuit any longer, but switch OFF the Gate of the freewheel FET very hardly. The only risk you are getting here is, that for some nanoseconds the freewheel diode becomes conductive. But this O.K. because then storage charge does not have time to build up.
You can make the circuit fast,
- by modifying the MOS Drive circuit and
- making the bipolars consuming some standby current of some mA or
- using RF transistors.
The benefit you have is also that the stability of the control is improved.
Greetings,
Matthias |
|
|
| Pafi |
mias!
This is not what I asked.
You are continously explaining the operation of the drawn circuit, but I asked you about what is not on the schematic! The usage / the rest of the circuit / the controll / the timing!
| quote: | B) When I put the
- input voltage
- in front of the totem pole stage
- to GND
I will switch OFF the FET in any case. |
You mean ON, don't you?
But the question was: when you do this (and when you apply the other state)?
("Where is the spade, father?
Next to the rake.
And where is the rake?
Next to the spade.
But where are they?
They are side by side.") |
|
|
| mias |
Hello Pafi,
| quote: | | You mean ON, don't you? |
Yes, I meant ON. The ON signal is coming in the phases
- where you have current flow in the direction of the body diode and in the phases
- with opposite current direction.| quote: | | In an amplifier this happens in every cycle at idle (and at low signal levels), but you mustn't allow switching off just because the current changes its polarity, FET have to stay ON till the next half-cycle! |
- I did not check this, but I promise you, there is no problem also when you have a change in the current direction. Here the FET is kept on being switched ON. The proposed MOS Drive circuit even does not realize that because it is only active when you make efforts to switch OFF the Gate. And then only in reverse mode. The circuit is just active for a couple of 100nano secs in the deadtime between the freewheel and the switching phase.
I do not have any sketch of the rest of the circuit. This was just a hobby investigation to understand the reverse recovery effect - not more.
As far as I know - there is no field of application up to now for this circuit up to now. It is not in use yet. Could be, it will become interesting. Especially for audio amplifiers where you have a high PWM frequency, it is worth to minimize the storage charge effects.
Maybe not nowadays, but MOSFETs are getting better and better, and then this circuit maybe quite interesting.
I have one question: How many Ampere are you switching when you use 200V FETs?
Greetings,
Matthias |
|
|
| Eva |
No particular effort is required to get 20ns transitions, they happen when drain current is high and MOSFET capacitances are charged very quickly as a result. If the crossover process is slowed down to, say, 200ns, to allow enough time for the circuit to act, then switching losses increase ten fold resulting in no overall gain.
Your circuit just doesn't have enough time to act. At high currents it should be able to turn on the MOSFET within one nanosecond or so. A "speed-up" capacitor in parallel with the diode could help a bit.
The solution that really works is precise timing of the gate drive signals with minimal drift and overlap.
Have you built and tested prototypes with real components? |
|
|
| mias |
Hello Eva,
The waveforms you are seeing are actual real measurements. No simulations or something like that.
The circuit can act pretty fast. All you need is to make the bipolar transistors fast, either with
- taking RF transistors
- or modifying the circuit so, that some mA collector current is flowing through the bipolars. Then they can react very fast.
And this is the best point, you can
- control the MOSFET in
- reverse mode
- with gate source voltage
- below the threshold voltage.
I showed this in the second pic I put in here (05-26-2008 02:24 PM).
This means as soon
- as the current in the freewheel FET
- reverses its direction
- the MOSFET is OFF.;)
| quote: | | The solution that really works is precise timing of the gate drive signals with minimal drift and overlap. |
My circuit does something very similar, believe me, but it just switches OFF the gate of the Freewheel FET, as soon as the the switching FET is switching ON. The time before the gate source voltage is kept to level nearby the threshold voltage.
Greetings,
Matthias |
|
|
|
