Hi Alexclaire
65 ohms for gate resitor is too much. Maximum practical value should be between 10-22 ohms. Try to add a simple resitor-capacitor-diode dead time at the input of the IR2113. At 65 ohms, rise time of your output stage will be too slow, and effiency reduced!
Fredos
www.d-amp.com
65 ohms for gate resitor is too much. Maximum practical value should be between 10-22 ohms. Try to add a simple resitor-capacitor-diode dead time at the input of the IR2113. At 65 ohms, rise time of your output stage will be too slow, and effiency reduced!
Fredos
www.d-amp.com
thanks Fredos,
your design is what i was thinking about...
dead time is the time when no fet is on, isn'it ?
a few ns i suppose
@250kHz which value would you choose for dead time ?
could you give me a value for R and C just to start with
1k 47pF would be a good start ? (with my simulator it gives 50ns delay)
thanks for your advice...
@12V bias, input voltage for IR2113 "1"logic level is 8V, so if i follow you dead time will be twice the time between the modulator 8V level output signal and 8V level input signal of IR2113...
hard to explain in english ;-) french would be better...
your design is what i was thinking about...
dead time is the time when no fet is on, isn'it ?
a few ns i suppose
@250kHz which value would you choose for dead time ?
could you give me a value for R and C just to start with
1k 47pF would be a good start ? (with my simulator it gives 50ns delay)
thanks for your advice...
@12V bias, input voltage for IR2113 "1"logic level is 8V, so if i follow you dead time will be twice the time between the modulator 8V level output signal and 8V level input signal of IR2113...
hard to explain in english ;-) french would be better...
I alexclaire
Si on y va en francais, les autres ne comprendrons pas!
I suggest you more to play with 470pF and lower value resistor, this will avoid noise pickup at the input of the IR2113. With 470pF, you will maybe play near 150-330ohms resistor. Yes, this will add dead time (time that both mosfet will be off and coil will work in freerunning mode, resonance). The value of the resistor will be dependend of the drive current capability of the stage before your IR2113 too..
Fredos
www.d-amp.com
Si on y va en francais, les autres ne comprendrons pas!
I suggest you more to play with 470pF and lower value resistor, this will avoid noise pickup at the input of the IR2113. With 470pF, you will maybe play near 150-330ohms resistor. Yes, this will add dead time (time that both mosfet will be off and coil will work in freerunning mode, resonance). The value of the resistor will be dependend of the drive current capability of the stage before your IR2113 too..
Fredos
www.d-amp.com
The 10Kw proto is still under developement. I have finish all 6 switching power supply and I'm working on the PFC. Result is impressive without PFC, +/- 416V output swing into 4 ohms, some saturation in output coil at this power, but bigger one is ordered. Bandwith is limited to 18Khz -3Db, due to dv/dt limit of the power stage at this power, effiency near 94%. My shop is heated by the heating plate I use to test it for the last 2 week! Sound good on speaker, but not the same quality as the 12KHVI, have to see why....I have not use the same power stage...slighty different, but front stage still the same! I have use IR2110 with transistor and baker clamp buffer, seem to change sound quality, maybe because this add more delay on the output stage. I will maybe try in a second prototype discret driver. Anyways, I donb have run it at more than +/- 260V on speaker, cone seem to try to go out of the basket!
Fredos
www.d-amp.com
PS for all who will ask why I do this project, it's for a heating floor test machine...But I whant to do one for audio, just for marketing show!
Fredos
www.d-amp.com
PS for all who will ask why I do this project, it's for a heating floor test machine...But I whant to do one for audio, just for marketing show!
Hi fredos,
very impressive !
All your HVI series amplifier have discrete power stage isn'it ?
2 * N channel Fets driven by opto ?
and bootstrap circuitry ?
could you tell me a little bit more because IR211* IC aren't very cheap in europe (7.5 Euro ~ 10 US dollars each) and very easy to burn ;-)
Thanks...
very impressive !
All your HVI series amplifier have discrete power stage isn'it ?
2 * N channel Fets driven by opto ?
and bootstrap circuitry ?
could you tell me a little bit more because IR211* IC aren't very cheap in europe (7.5 Euro ~ 10 US dollars each) and very easy to burn ;-)
Thanks...
The second RC section is to slow down the transient response of the amplifier. Too fast transient response make amplifier unsatble under certain load. Mosfet driver of the power stage of the HVI are made with optocoupler and IC 9A mosfet drivers, with 8A 35ns bootstrap diode. Basic way, but mosfet are direct drive with only ferrite bead on the gate, and dead time are controled by electronic after the LM319 stage. Only current at transition are sensed and dead time are manipulate with this measured current. Dead time are always optimal at any power/load/frequency. Current sense are done by 0.005 ohms pcb track resistor. Main output current sensing are done on the switching side by a current transformer and maximum power versus time are controled by a timer there and are dependent of output switch temperature. Feedback are direct without compensation like a analogue amplifier. So all current lose are corrected by the overall feedback. From 1000 ohms to 2 ohms frequency respond is flat to 22 000Khz, and power bouble form 8 to 4 ohms and at 2 ohms dynamic power double, but long term power is limited to 25% of RMS power into 2 ohms.
The operation of the 10 Kw version is nearly the same except a different power stage.
Fredos
www.d-amp.com
The operation of the 10 Kw version is nearly the same except a different power stage.
Fredos
www.d-amp.com
i am using 220pf an 1k res at each input of 319 and it was quite stable ,so i decided to change only the one at - input to 2k7 in order to hear the difference but the amp became unstable after 2 seconds needed to change the offset compensation to become stable again .
Quite impressive the description of your new power amp.How about power supply?Regulated or not?Will you sync?What frequency of power amp?Is it bridge single power or half bridge split power?
Quite impressive the description of your new power amp.How about power supply?Regulated or not?Will you sync?What frequency of power amp?Is it bridge single power or half bridge split power?
For the HVI: Power supply unregulated: more efficient and load regulation is keeped within 10%. Only load detection to switch between normal/hickup operation. Overall amplifier only take 25 watts at idle. Power stage is floating ground referenced full bridge, single supply. Power supply is at 62.5Khz and power amp run at 500Khz.
On the 10Kw beast, regulation loop is done at the PFC and include switching PSU stage in the loop. Regulation is inside 1% from 0-100% load. Power stage is floating ground referenced modified full bridge, single supply at 450V. Power supply is at 62.5Khz, PFC at 125Khz and power amp at 250Khz.
Fredos
www.d-amp.com
On the 10Kw beast, regulation loop is done at the PFC and include switching PSU stage in the loop. Regulation is inside 1% from 0-100% load. Power stage is floating ground referenced modified full bridge, single supply at 450V. Power supply is at 62.5Khz, PFC at 125Khz and power amp at 250Khz.
Fredos
www.d-amp.com
500khz!Its very hi frequency! i imagine how many hard problems to make in such hi freq.
I saw one qsc pl9.0 working like floating ground.not so sure it uses pfc,but the freq it is not to hi.not regulated ps .lm393 on the comparator but i dont know if its class bd or ad,didnt check.It uses optocoupler .Does not sound so impressive maybe becuse of the subs and enclousure.Not sure.It has a complex way to limit pwm.
I saw one qsc pl9.0 working like floating ground.not so sure it uses pfc,but the freq it is not to hi.not regulated ps .lm393 on the comparator but i dont know if its class bd or ad,didnt check.It uses optocoupler .Does not sound so impressive maybe becuse of the subs and enclousure.Not sure.It has a complex way to limit pwm.
PL9 are basic linear amplifier. It use current cell to share power to the huge output transistor. It have 8 cell (2 in // on each side of the full bridge). Current cell are used as a buffer to the voltage amplifier. Switching use rudimentary 1 stage flyback PFC converter, not very efficient. Nothing PWM in amplifier stage.
Fredos
www.d-amp.com
Fredos
www.d-amp.com
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.