| Joules |
| What happens when class D clipp's? What does it sound like? as compared to, say, a class A/B ? |
|
|
| SSassen |
Depends on the topology used, the Hypex UcD modules sound very much like a standard class A/B amplifier. But really, why would that be a concern? Will you be running a class-D amplifier near its maximum rating? If so, you need a more powerful one perhaps :xeye:
Best regards,
Sander Sassen
http://www.hardwareanalysis.com |
|
|
| phase_accurate |
This will mostly depend on the feedback topology used. Some additional soft-clipping measures can be taken as well.
If no soft-clipping tricks are used the following applies:
The least nasty (i.e. cleanest) clipping is from non-NFB topologies closely followed by first order loops with a high dominant pole freqency.
Higher order loops have the habit to "stick to the rails" when driven into clipping.
Regards
Charles |
|
|
| nitrate |
Greet's
I was running my amp into clipping for a prolonged time over the new year ( celebrations and all that ). The amp is a home built self osc classD of my own design. The first thing that goes is the bass, this is followed by audiable clipping across the rest of the spectrum that sounds quite soft in comparison to A/B clipping. I would say the amp sounds like it backs down during clipping. I would not describe it as a harsh or damaging clip sound. I was rather drunk but it can't have sounded that harsh as i let it carry on like that for a large part of the early morning celebrations, i wouldn't delibratly harm my baby LOL
I can't speak for the other amps on here but i think the clipping is certaintly no worse than A/B. I gather you are worried about the amp self destructing at clip or somthing. I know i was, but after a few partys my confidence has grown and i now know it is rarther hard to damage these amps. The only time mine is destroyed nowadays is when i'm hot pluging and removing components from the breadboard.
Regards
Mad.P
P.S The amp was running 80RMS into 8R when clipping. Lasted all night! |
|
|
| SSassen |
Nitrate,
| quote: | | i now know it is rarther hard to damage these amps. |
Try short circuiting the output, your output MOSFETs and fuses will blow faster than you can blink an eye. For added effect do this at a high output rating, i.e. driving the amplifier and connected loudspeaker at considerable volume, now you will see sparks fly :D
Best regards,
Sander Sassen
http://www.hardwareanalysis.com |
|
|
| phase_accurate |
You can even run them into clipping at quite high-frequencies for prolonged periods. Now try this with the average AB amp out there !
Regards
Charles |
|
|
| Workhorse |
Hi Sander,
Please have a mercy on Mad.P's amp..donot direct him to short circuit his amp...donot rule his mind :D :D :D :D :D ;)
K a n w a r |
|
|
| Workhorse |
| quote: | Originally posted by phase_accurate
You can even run them into clipping at quite high-frequencies for prolonged periods. Now try this with the average AB amp out there !
Regards
Charles |
Dear Charles Leymann,
Nothing happened only harshness in Treble was encountered....
K a n w a r |
|
|
| phase_accurate |
Hi Kanwar
How is your amp BTW ? Anything fried so far ?
Regards
Charles |
|
|
| Workhorse |
Nothing Fried in the amp...but it fried a 20MFD cap connected as a shunt to check out capacitive load drivabilty...that all....
One question...In my POST filter feedback amp when no load connected..i.e. open circuit ..does the feedback automatically take cares of Q of Filter.... when no load is connected...I have cliiped the amp with open circuit and nothing happened...
amazed...
K a n w a r |
|
|
| SSassen |
Workhorse,
| quote: | | Nothing Fried in the amp... |
Try the same suggestion I gave Nitrate, I'm sure that'll have some effect, and do post pictures afterwards :D
Best regards,
Sander Sassen
http://www.hardwareanalysis.com |
|
|
| nitrate |
LOL,
Don't worry folks i've already tried that one! Not by intent, i was running the amp into a dummy load with wobbly croc clips on the output terminals, of course they wobbled toward each other then the power supply decoupling resistors fried open circuit ( saved most of the amp ). Even the smallest signal on the output instantly blows these amps due to the very high damping factor. The amp just stuffs more and more power into the output terminals to try and get it to reproduce the input signal ( it does on mine anyway ). You could say our amps are TOO good for their own good LOL And yes, i take back my statement of these amps being almost undestructible, they are extremly delicate when it comes to short circuits or capacitive influences in the modulator section ( your fingers for instance ). However i will say that under normal operation the classD amplifier is very robust and almost impossible to damage by overdriving or overloading ( as long as its not a sudden short ).
Have fun all,
Mad.P |
|
|
| Workhorse |
Hi Sander,
Nothing would happen because i have a short circuit protection onboard.....:D :D :D :D ;) hahahahaaaaa
you cannot rule my senses.....:angel:
K a n w a r |
|
|
| SSassen |
Nitrate,
Indeed, that's the culprit with these amplifiers, they'll continue to try and drive even a 0-ohm load with resulting excessive currents that'll either blow the fuses, or the MOSFETs. I've been using current restricting resistors as well, and have seen them glow red-hot after a shortcircuit. As for fuses, I'm currently using 6.3A Slow on my PSU and they're too slow, I lost two pairs of MOSFETs already due to crocodile clips touching each other and a woofer shortcircuiting its voice-coil.
Anybody thought up a effective short-circuit protection already? I like the way Hypex does it, but I'm a little reluctant to try and find out of it works, as that'll mean shorting a perfectly fine UcD180 under load. Any ideas?
Best regards,
Sander Sassen
http://www.hardwareanalysis.com |
|
|
| phase_accurate |
| quote: | | One question...In my POST filter feedback amp when no load connected..i.e. open circuit ..does the feedback automatically take cares of Q of Filter.... when no load is connected. |
The most accurate answer is: It depends !
This simple feedback topology is still prone to instability with critical loads. You seem to be on the safe side when it survived 20 uF of capacitive load and clipping into open output.
Regards
Charles
Edit: The most interesting question: How does it sound ? |
|
|
| Workhorse |
Thanks Charles....
I have tested both pre & post filter feedback topologies under your guidance....During LIVE Vocal singing midrage test..pre-Filter feedback has more tranperancy with defining the image of vocals...I dont know it why...maybe you have some thoughts on this too....
K a n w a r |
|
|
| SSassen |
Workhorse,
Your amplifier is that a self-oscillating design? Or one that has a external oscillator? And how did you implement the short circuit protection that's something I'd be very interested in, as I'm currently researching that bit, I got the DC protection figured out already.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com |
|
|
| Workhorse |
Hi Sander,
Mine is External Clocked..3 Level PWM SWFQ@270KHZ ..2KW@2Ohms.....
I have implemented a resistor in series with lower rail which is connected to FET source terminal of Bridge amp for sensing and activating 555 timer for 5 seconds for muting the input signal of amplifier ....
K a n w a r |
|
|
| SSassen |
Thanks Workhorse,
I believe I've seen pictures of your amplifier in a different thread, it is for PA use mainly correct?
Best regards,
Sander Sassen
http://www.hardwareanalysis.com |
|
|
| Joules |
Excuse me while I Highjack this thread :tongue: :tongue: :D
I have a few CD's with some very high level deeeep bass and can give any bass system fitts. I haven't cliped my pair of 12 Kw ICE amps yet but was just querious. |
|
|
| Workhorse |
Hi Sander,
Yeah we manufacture pro-amps for addressing large public gatherings...but some folks also use our amps in there home theatre systems to power up their subs to blast their houses and made their neighbours into run.....
The pictures you have seen must be some of our obsolete Class-AB amps....which are no longer in productions.....Our current range include Class-D and Class-AB NVMOS only.....
K a n w a r:D |
|
|
| Workhorse |
| quote: | Originally posted by Joules
I haven't clipp my pair of 12 Kw ICE amps yet but was just querious. |
Do 12KW ICE amps really exist...do you have a weblink ... |
|
|
| nitrate |
LOL,
Sassen, Sounds like you have the hands on approach like me :smash:
If you come across a protection circuit that looks like it could actually work without intorducing more problems at high power then please let me know! I've spent many nights thinking of ways to protect things but all attempts have failed. I've gone through loads of pairs of FETs all with the same cause of death. Over current! either from speaker shorts or silly inputs ( usally caused by unplugging the RCA inputs live ).
I must say however, under normall operating conditions i've run the amp into a 200W 4R resistor clipping so hard it just shows a square wave on the scope at 10KHz at PK-PK around 60V ( over 100W RMS )for more than 10mins. Not only did the amp survive but it took the Wee Wee by staying cool!
Regards
Mad.P |
|
|
| Workhorse |
Thanks, for correct info, the highest we have is 5KW for driving subs only...
K a n w a r |
|
|
| SSassen |
Nitrate,
| quote: | | Sassen, Sounds like you have the hands on approach like me |
Well, yes, I simulate/calculate things first, but then just don't sit back and marvel at my creation but start thinking of ways how during use it could be made to fail, and test these out on the prototype. Short-circuiting outputs is a far too common occurance to simply ignore for example. I do believe simulation and calculus are a great way to get a amplifier designed and up and running, the proof of the pudding for me is still in the actual measurements of a prototype and of course a close examination of under what conditions it might fail.
If I come across anything that's working and doesn't interfere with the amplifier too much I'll be sure to let you know. Currently I'm using a LM2941 regulator that has a TTL compatible on/off pin. I use sense resistors in the power supply lines and a simple transistor stage to boost the output to logic levels. The LM2941 powers the driver IC (IR2011) and hence when a overcurrent is detected it simply switches off the power to the driver IC. DC protection is implemented by measuring the output and having two relays in the power supply switching off both rails. Similar to Hypex, but designed differently.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com |
|
|
| phase_accurate |
Hi Kanwar
I will have a look at the e-mails again to see how the NFB actually was dimensioned.
Regards
Charles |
|
|
| zilog |
Sander,
Do you implement any power supply voltage guard or anything similar, for instance to ensure proper high side driver capacitor charging before switching starts?
I am interested in seeing your schematic for whatever you have implemented as I also have come up with the idea of using a 555.
/Daniel |
|
|
| raintalk |
| Anybody have a circuit for a clipping indicator/limiter for use with Class-D? |
|
|
| phase_accurate |
One way to detect clipping is via a retriggerable monoflop. The output of the comparator retriggers it every time when the signal is going from low to high (or vice-versa depending on implementation). So it will stay in the metastable state if the time-constant is set accordingly. A soon as the transistions stop (which will happen during clipping) it will return to the stable state which can then be displayed. Maybe a second monoflop should be used to turn on the indicator long enough to be visually perceived.
The output of the first monoflop can also be used to control a VCA (after soem processing of course).
Regards
Charles |
|
|
| Workhorse |
Hi everyone,
I have seen something else in another reputed Class-D amp....That amp restricts the input of amplifier with VCA before its output voltage reaches the threshold of clipping..thereby it never clips in reality....Thus the output never touches its power rails..its voltage output remains just below 2 volts of supply rails.
regards,
K a n w a r |
|
|
| kartino |
Hi,
I still thinking about clipping in class-D too.
Let say that voltage rail is 50V, hence the class-AB will clip at 50V minus voltage drops.
Can the output voltage of class-d reach as high as class-AB before clipping occur? (Consider to the pulse width, t-on/t-off time, dead time, length of mosfet on-off etc.)
Thanks
kartino |
|
|
|
