Hello
How about Silicon Labs Class D Driver IC SI8241 / SI 8244 ?
isolated drivers, voltage up to ±750 V input-to-output and output-to-output isolation,extremely low propagation delays enable faster modulation frequencies for an enhanced audio experience
Why this IC not becoming popular in DIY and Class D amplifier ?
Everywhere all known commercial, DIY and chinese Class D Designs / Amplifier using IR / Infineon or TI Chips
How about Silicon Labs Class D Driver IC SI8241 / SI 8244 ?
isolated drivers, voltage up to ±750 V input-to-output and output-to-output isolation,extremely low propagation delays enable faster modulation frequencies for an enhanced audio experience
Why this IC not becoming popular in DIY and Class D amplifier ?
Everywhere all known commercial, DIY and chinese Class D Designs / Amplifier using IR / Infineon or TI Chips
I am no specialist in these matters but it seems SI824x may replace IR2110 or IR2113 as high-/low-side drivers.
IRS2092S includes a PWM-modulator with the high-/low-side drivers such that a class D amplifier can be implemented by adding power switches.
Thus, as the delicate PWM-modulator is missing, SI824x will only do a part of the job. TC2000/2001 may be used as PWM-modulator but then you end up with a PWM modulator IC, a high-/low-side driver IC and separate power switches. TC2000/2001 is better used with a power bridge like STA508/516 including the driver circuit. And, you get all in a monolithic class D amplifier chip.
For audio use, SI824x seems obsolete just as IR2110 and IR2113.
Once more, I'm no expert in these matters.
IRS2092S includes a PWM-modulator with the high-/low-side drivers such that a class D amplifier can be implemented by adding power switches.
Thus, as the delicate PWM-modulator is missing, SI824x will only do a part of the job. TC2000/2001 may be used as PWM-modulator but then you end up with a PWM modulator IC, a high-/low-side driver IC and separate power switches. TC2000/2001 is better used with a power bridge like STA508/516 including the driver circuit. And, you get all in a monolithic class D amplifier chip.
For audio use, SI824x seems obsolete just as IR2110 and IR2113.
Once more, I'm no expert in these matters.
Last edited:
This part you note is a very, very primitive building block of a class D amp--it's maybe 5% of the total design effort needed. It's very hard to make a discrete class D amp and beat what TI is achieving with their 300W closed loop amps.
Plus, the flexibility of the modern class D turn-key parts from TI is huge: You can decide to make a single 600W amp, or stereo 300W or quad 125W.
I understand, but for HIGH POWER Class D > 1000 W Rail > +/- 95 V
discrete design Half / Full bridge with IR2110 or IRS20957 with PWM modulator like Behringer and many competitor does is required
I think in this case Silicon Labs Class D Driver IC SI8241 / SI 8244 coukd be good substitute Vs IR 2110 / 2113 FAN7392 / IRS2092 / IR20957
datasheet offer much benefits compare IR / Infineon IC driver but cant see this IC in DIY and fiished commercial audio product
all DIY and comercial Half or Fullbridge designs and Chinese PA factory using IR
Are you sure this is all comparing the same thing? For example, Behringer specs max power in the NXS6000D with no qualifier on THD. A TI TPA3255 specifies continuous power. Have you seen any real world measurements on a Behringer class D?
This IC isn't getting the attention it deserves. 🙂
Interesting post. I'm putting a parts list together for a new design I'm playing with, I'm interested in this
The SI 8244 seems to be have better specifications for high end amplifiers aimed at designs with low noise immunity goals and high current buffer drivers for
Kw designs.
pro:
- High-precision linear programmable dead-time generator 0.4 ns to 1 µs !
- 0.5 A peak output (Si8241), 4.0 A peak output (Si8244)
- Isolated driver, (proprietary isolation technology) not optically based.
- extremely low propagation delays. 60nS
- Typical shutdown command pin (Shutdown Time from disable true) time 60ns vs 500nS for IRS20957
- Cost ~$3
Lack of on-board current protection trips. (however the use of a IR25750LTRPBF can make up for this) also this chip is ideal for a modern design where micro-controller supervision is preferred.
I'm ordering a few to test with.
Last edited:
Si824x is obsolete.....they have now...https://www.silabs.com/documents/public/data-sheets/Si823x.pdf 🙂
The prop delay is 25ns, not 60ns (that's the worst case, ie at max temperature), and more importantly the shortest pulse width and pulse distortion figures are good at 10ns and 1ns (typical).
4A peak output is not very special in the world of fast gate drivers, you need high currents to switch fast and accurately in the time-domain. I know of 9A drivers (you can implement a small class-D amp just using MOSFET drivers in fact!)
The very wide range of the programmable deadtime is definitely a good feature. They don't actually quantify the phrase "high-precision" alas, so it basically has no meaning - basically any adjective without a numerical value to back it up is marketing, not specification 🙂
Let us know how you get on with it!
4A peak output is not very special in the world of fast gate drivers, you need high currents to switch fast and accurately in the time-domain. I know of 9A drivers (you can implement a small class-D amp just using MOSFET drivers in fact!)
The very wide range of the programmable deadtime is definitely a good feature. They don't actually quantify the phrase "high-precision" alas, so it basically has no meaning - basically any adjective without a numerical value to back it up is marketing, not specification 🙂
Let us know how you get on with it!
How can they be obsolete if I made 500 pc order from SiliconLabs one month ago?hm
Quite often edge rates need to be reduced for EMI in commercial products so the high speed gate driver is made slower using series resistance etc. The EMI issues can be overcome by using lower charge and inductance parts that have less overshoot (EG GaN). These parts don't need massive gate drive current and are usually limited in how fast they switch by the driver slew rate and gate loop inductance rather than gate charge. Check out this part for fast switching (0.9nS):
https://www.psemi.com/pdf/datasheets/pe29102ds.pdf
Would be capable of 2 kW into 4 ohm BTL amp.
The disable (shoutdown) pin is armed with a positive signal referred to ground, the 2110-2113 referred to the-V, another pros is a patent complete isolation and dead-time generator, in iraudamp design (NO irs2092, irs20957) there is no dead time.
Do you think it could be connected as shown?
Do you think it could be connected as shown?
Attachments
Great stuff! I see the next successor to the Ncore class-d amplifier is using it! The specification on this particular IC looks promising, hope to build something using it over the beloved IRS20957S
Attachments
I have used the SI 8244 with good success.
I have chosen IRS2092 in all my current designs, as it includes a OTA amp which can be used as a comparator, and because of the built in current limiter.
For higher power the SI chip has a big advantage of high output current of 4A and higher voltage .... for e.g. IRF4227 you need to use buffers with the IRS2092.
I have chosen IRS2092 in all my current designs, as it includes a OTA amp which can be used as a comparator, and because of the built in current limiter.
For higher power the SI chip has a big advantage of high output current of 4A and higher voltage .... for e.g. IRF4227 you need to use buffers with the IRS2092.
Nothing to do with power. Both chips are equally good.However the SI 8244 is more of a low level chip for amplifier designs pushing beyond normal specifications, attractions are: incredible dead-time limit for ultra low THD figures, silicon isolated barrier between the driver and signal front-end, protection(s) is now in the hands of of the designer where 1us shutdown figures maybe achieved compared to the ~250us propagation shutdown command.
It will probably be really difficult to tell the difference in playback mode, expect though when it comes to defense, failure and recovery implementation.
IRS2092 is limited to +-100V, and Io+:1A, Io-:1,2A ..... This for sure limits the output power at a given Impedance and, the type of FETs you can drive directly
Although true, isn't practical reasoning, high power designs are between 75V to 90V and almost always limited to the power drain to source break down. of which 200V is almost the standard yield.