| Steven |
Philips released its UCD application note.
http://www.semiconductors.philips.c...s/UM10155_1.pdf
The UcD concept was patented by Philips. However, the use of the IP is granted (without a license fee) to third parties that use the enabling semiconductor components from Philips Semiconductors in their application. |
|
|
| IVX |
| Fig 1.General view -"Philips *** D amp ".. Funny, but If Philips would be my own, I would dismiss the photographer! :cool: |
|
|
| djQUAN |
| a good starting point for a DIY class D amp..... just printed it and gonna have a read....... |
|
|
| classd4sure |
>120dB SNR?? ... and with this single ended version? hmmmmmm.
It's an OK read, but you'll learn more from going to the hypex website and reading an actual paper written on it by the guy who actually knows something of it, and much more yet from the posts on this forum if you go back far enough.
Did they give the "most important" information or the most basic? According to Bruno's comment the last time this was brought up.... this is all they've got to give on it. Funny. I guess that explains them having "serious problems implementing it" (so a little bird told me)... guess they should have stayed on his good side?
Funniest part:
"Again, any changes are left the responsibility of the experimenter, but assistance from
Philips will be available through the local sales representatives."
What part of that do you think gets removed upon the next revision of the datasheet.
Ivan, about the pic.. it's your typical microfilm quality, they must have stolen it when he had his back turned :D |
|
|
| djQUAN |
I've done some looking around in it and didn't find anything about input sensitivity or gain......
although the simple schematic appeals to me..:D |
|
|
| classd4sure |
Well, there's alot of reading you can do in the areas I mentioned, but if you really want to get your feet wet and jump in head first, just build this sucker and use as advised. I wouldnt' expect that performance from it though.
This cheapness is likely to be the reason why Philips can't make use of it! So if you want better... you're going to have to dig up a little more on it, but it's all out there.
EDit: You're not going to want to judge a Hypex version based on this sucker though! |
|
|
| djQUAN |
well, I'm not really looking for the "best class D that money could buy" thing but something quick and dirty to drive a pair of subs. probably build this first as is and work up to a higher power version when I get a feel of things.
edit also:if you got a link for a schematic with more power and better performance as is then I'm all ears (or eyes.:D ) |
|
|
| fokker |
that's an interesting link. appreciate it.
i am still struggling with mine - which more and more looks like a mosfet driver issue but I will report back when I get it fixed.
classd4sure, what is the "most important" information? I think a good circuit is one that is first of all robust. If something is that quirky, there usually is something foundamentally wrong with the circuit. |
|
|
| classd4sure |
| quote: | Originally posted by fokker
that's an interesting link. appreciate it.
i am still struggling with mine - which more and more looks like a mosfet driver issue but I will report back when I get it fixed.
classd4sure, what is the "most important" information? I think a good circuit is one that is first of all robust. If something is that quirky, there usually is something foundamentally wrong with the circuit. |
I don't know I tend to think alot is equally important oppositly than they've pointed out. You don't have to worry about it though, because all that stuff doesn't appear in their version of it :)
You're using a driver IC right? Maybe post some experiences/schematics to your thread and we can help you out with whatever issues?
Cheers,
Chris |
|
|
| janneman |
Well, a lot of comments I see here are of the variety "it's simple, therefore it can't be good". Does anybody actually understand how it works?
Jan Didden |
|
|
| IVX |
| quote: | Originally posted by janneman
Well, a lot of comments I see here are of the variety "it's simple, therefore it can't be good". Does anybody actually understand how it works?
Jan Didden |
It's work pretty well, but Hypex UcD a little bit better, because little more complicated. |
|
|
| classd4sure |
| Did you measure >120dB SNR Ivan? |
|
|
| IVX |
| quote: | Originally posted by classd4sure
Did you measure >120dB SNR Ivan? |
No i didn't build simplified UcD from this datasheet yet. Actually, why not -120db? BTW, UcD400 in the my setup, gave about -110-115db noise floor at 1KHz@150w/8ohm, however idle noise floor was much lower, so Philips can recalculate S/N as S/N(idle), it will much more cool even :) |
|
|
| classd4sure |
I have only ever seen it rated as being >110dB.
I just doubt their single ended version would do >120dB, that's all. :) |
|
|
| janneman |
| quote: | Originally posted by IVX
It's work pretty well, but Hypex UcD a little bit better, because little more complicated. |
Yes, that's logical. It is more complex therefore it must work better. Maybe I can add anothe 10 components to make it more complicated so it even works more better.
Do you actually know how this circuit works?
Jan Didden |
|
|
| classd4sure |
| quote: | Originally posted by janneman
Yes, that's logical. It is more complex therefore it must work better. Maybe I can add anothe 10 components to make it more complicated so it even works more better.
Jan Didden |
Guess I should not have deleted my prior post.
Here it roughly is again.
Bruno's designs are probably the most elegant I've ever seen. As I'm sure you know patents rarely include "everything" towards a preferred implementation. This is such a case.
If you just go to the Hypex website and read their AES paper on UcD you'll instantly spot one obvious difference... differential feedback. There's yet another improvement which just makes it a more robust circuit, and is needed at higher power, that often appears on this forum. There's a few more things that set it apart from this version as well that we don't necessarily know about. Philips doesn't have any of this, and I guess they'd rather push the digital envelope with far more complex circuits than put the little bit of effort required into fixing this one up a little.
I got word from someone that Philips basically gave up on this version of UcD because they had alot of problems in implementing it. Why is that when Hypex has no such issues, and neither does the vast majority of their customers... why only Philips .... who own the patent? Because, all they have on it is the "as seen on patent" version. Several have managed to build diy versions based on designs floating around here, that are better then what you see in this datasheet by philips in some ways. They did it without even the use of a scope, which is absurd. I'm one of the first to have done so too. Ivan was also one of the first to be playing around with his own discrete versions of it if I'm not mistaken.
It's not a question of complexity being better. They didn't share this because it's their best amp going, they shared it because they gave up on it, so why not use it to push their transistors instead.. |
|
|
| Bonsai |
I know the guy at Philips who built this and I've seen the amp (they made about 30 PCB's). I never heard it but another guy at Philips (an applications engineer) compared it to his Arcam Alpha (DIVa85 I believe). He told me it was pretty impressive - good sound.
Layout absolutely critical especially around Mosfets (circuit is optimised for the mosfets shown, so do not change them unless you are in a postion to do some tweaking) and output inductor. All traces short as possible. In fact, trhe mosfets on the board are back to back with the hesatsink in between them in ordeer to get the traces as short as possbile.
I guess this circuit is not as easy to implement as an IC based design, but if you are a manufacturer looking for a high powered design, this makes a lot of sense because you can get the repeatability from board to board. Nevertheless, its still a fun circuit I imagine to play with.
for me personally, I stick to standard linear designs!
bonsai |
|
|
| selfservice |
anyone can explay how this design function?
i've ipotized something but i'm not sure |
|
|
| classd4sure |
Hi,
Go to www.hypex.nl and click on "technology" link. That's basically the AES paper on it.
For further reading you can dig up the patent, which may explain the basic funtion of it in greater detail.
Hope that helps. |
|
|
| lenz |
| quote: | | why only Philips .... who own the patent? |
Well, I think : it`s Philips.:cool: :cool: |
|
|
| classd4sure |
| Right, so you'd think they of all people could pull it off? Too many cooks in the kitchen perhaps? |
|
|
| selfservice |
i've simulated the circuit with pspice
the result is a very important residue of high frequency on the output load and the sinus voltage in output is shifted down, so i've -40V for the peak negative voltage and only +30V for the positive peak voltage.
i'dont know if this is because i've used some equivalent models instead of philips suggested components or i've do something wrong in the circuit or unfortunately the circuit is really wrong.
the only sure thing however is to phisicaly build the circuit and attach a scope on it :D |
|
|
| classd4sure |
| Tried making R14 and R15 = ? |
|
|
| selfservice |
| yes, no change... i'm going to try lower values |
|
|
| classd4sure |
| May I see your simulation? |
|
|
| nFORCE |
Philips Semiconductors change to NXP
www.nxp.com |
|
|
| jacco vermeulen |
NXP, what-if-you-could-sell-it.
(80,1% : Kohlberg Kravis Roberts (KKR), Silver Lake, AlpInvest Partners, Bain Capital en Apax. )
|
|
|
| RX5 |
| quote: | Originally posted by classd4sure
Well, there's alot of reading you can do in the areas I mentioned, but if you really want to get your feet wet and jump in head first, just build this sucker and use as advised. I wouldnt' expect that performance from it though.
>>>>>>>>>>>This cheapness is likely to be the reason why Philips can't make use of it! So if you want better... you're going to have to dig up a little more on it, but it's all out there.<<<<<<<<<<<<<<
EDit: You're not going to want to judge a Hypex version based on this sucker though! |
yep.... very true indeed :D but if circuit "tuned" properly, AMP is really GOOD :) all info is OUT there just waiting to be grabbed and explored....:smash: |
|
|
| rogs |
The overcurrent protection circuit for this application is not included on the main circuit diagram on page 4, nor are the components included on the demo pcb. Probably not essential for simple testing, but useful to try.
The additional circuitry is shown on page 9 --- but I can't see how it works. It looks to me as if the zener diodes D1 and D2 are drawn in reverse - surely not the same polarity for both positive and negative halves?
Or am I missing something? |
|
|
| RX5 |
| quote: | Originally posted by Steven
Philips released its UCD application note.
http://www.semiconductors.philips.c...s/UM10155_1.pdf
The UcD concept was patented by Philips. However, the use of the IP is granted (without a license fee) to third parties that use the enabling semiconductor components from Philips Semiconductors in their application. |
waaaaaaaaaaaaaaaahhhh .. how the heck did I miss this thread/PDF? hehehe maybe Ive been Sooooooooooooo busy .... to read!!
:cannotbe: :cannotbe: :D anyways, hmmmmmm If anything interesting comes up, gonna re DO PCB layout again... 3rd time... :smash: |
|
|
| IVX |
| One my buddy already tried PCB, captured from the UM10155_1.pdf, he said -pretty stable, but have not measured (THD) so far. |
|
|
| fokker |
| I thought repeat claims have been made here that Phillips couldn't make it work. is that true? |
|
|
| RX5 |
| quote: | Originally posted by IVX
One my buddy already tried PCB, captured from the UM10155_1.pdf, he said -pretty stable, but have not measured (THD) so far. |
yup...it WILL work.... :) but it certainly needs some minor improvements.. based on what classd4sure has posted a few years/months ago.... |
|
|
| RX5 |
| so ..... any OTHERS here actually DID the circuit?? :D |
|
|
| fokker |
| quote: | Originally posted by RX5
so ..... any OTHERS here actually DID the circuit?? :D |
why would anyone build something that supposedly doesn't work, per the experts?
actually, why would phillips send out evaluation boards that don't dowk, if the experts are right?
or maybe the experts who insisted that the design doesn't work are really clueless on the subject matter? |
|
|
| alfsch |
i made my "ucd400"-version , basically same circuit.
works really fine, up to +- 40v.
(at higher voltage a lot of problems) |
|
|
| classd4sure |
| quote: | Originally posted by fokker
why would anyone build something that supposedly doesn't work, per the experts?
actually, why would phillips send out evaluation boards that don't dowk, if the experts are right?
or maybe the experts who insisted that the design doesn't work are really clueless on the subject matter? |
:king: :joker: :smirk: :violin: |
|
|
| fokker |
| quote: | Originally posted by alfsch
i made my "ucd400"-version , basically same circuit.
works really fine, up to +- 40v.
(at higher voltage a lot of problems) |
sounds like a slap or two on the face of the experts who claim otherwise. |
|
|
| RX5 |
| quote: | Originally posted by alfsch
i made my "ucd400"-version , basically same circuit.
works really fine, up to +- 40v.
(at higher voltage a lot of problems) |
HOW would u say its 400??
:bigeyes:
if your supply is only +-/40V :D
what problems btw?? ;) |
|
|
| alfsch |
well, it should work at +-80v ...
the real problem is the mosfet: i have no source for the real "nice" fets, like philips PHP28NQ15 ...so i used irf640n. and here are the problems !
Cg too much, back diode slow >> too much idle current + too slow switching at high supply.
maybe i try with harder gate-drive, but i made a smd-board, so not much space for bigger driver. or i get "better" mosfets... |
|
|
| audiomanics |
Unit shuts down at 68V on either rail..
75V is absolute max for this UcD400
40-0-40 is fine, for the transformer that is..
Kees |
|
|
| classd4sure |
| Even with a better mosfet you'd want to upgrade the drivers somewhat at these levels. |
|
|
| alfsch |
ok, right, but you never know, if you dont try...
or with other words: in swcad simu running fine, (nobody told me, that the mosfet models are sh.. , only "guru" bruno wrote, he dont use simu, because real life is too different) , real amp fine , but at 40v and up, real amp is no more like simu !
in test: without mosfet, but with 10nf load for the driver switching is ok.
so i wonder: irf640 at +-70v shows much more "nf" Cg !!??
or i have a bad magnetic effect, that slows down the driver...
you could see it here:
d-amp |
|
|
| classd4sure |
With this it goes right back to the loading of the comparator stage/current source.
Output current gets higher and higher, all transistors suffering miller effect (down to the current mirrors) get weighed down more and more by it.
Any point along the chain you can minimize this effect should likely be taken.
There's also all kinds of other aspects that come into greater play because of all the higher currents running around, they're likely to require snubbing/control at higher power.
The use of the cascaded output stage on the comparator does alot to even things out and make it less susceptible to the miller current slewing around. It will also help ensure equal temperature of the mirror transistors. That's the reason for them.
Even fokkers basic fumbling with gate driver IC's is only good up to a certain power level before you have to know what you're doing.
You can still use the simulator as a baseline, if you design drivers with the right properties it's quick to check basic operation with a simulator, final values will have to come from real world teaking.
There's other factors why this amp isn't as good as even the more basic DIY versions.... which are more Hypex like in implementation using fully differential feedback etc.
It would have been cool to have heard more about your efforts with this, is it still a work in progress?
Ahh yes, I had seen that, but did not follow because it's not my language.
|
|
|
| RX5 |
hmmmm... im running MY diy UCD at +/- 55V .. and only IRF540.... :)
no problem whatsoever... |
|
|
| fokker |
| quote: | Originally posted by RX5
hmmmm... im running MY diy UCD at +/- 55V .. and only IRF540.... :)
no problem whatsoever... |
that's interesting. would you mind sharing with us your schematic and component selection?
Thanks. |
|
|
| RX5 |
| quote: | Originally posted by fokker
that's interesting. would you mind sharing with us your schematic and component selection?
Thanks. |
dear fokker (and others)
months ago, I think you have seen my earlier post(s), about my D-amp, the UCD way... I 1st started out a circuit something VERY SIMILAR (note something VERY) to the philips app note....It worked even with BAD component selections, then after some parts swapping and some circuit simulation and some PMs from experts and the inventor himself, I tried it with the cascode type / 2 transistor current source...
ALL did it in a few months time, swapping/simulating/guidance from experts.. until I finally settled with optimized components...
I AM ONLY using IRF540 because it is the best part that I could find here in our place... got no extra money to order those "nicer" parts.. too bad.. but at least, THIS PROVES that even with "less than nicer parts", you can stil have quality amp.... I have junked my 100W AB amp, in favor of my diy UCD.... :)
Classd4sure(Chris) has ALREADY posted his running schematics and with some changing of parts from sim to real world, IT HAS WORKED! I have NOT deviated from that circuit since, just tweaked values... I just added some extra features (like over voltage protection)... got it here in THESE forum... and the over current protection?? I had an idea how to but was not sure how to implement it....had to do some simulations again(and some help from Bruno himself, thanks bruno)... so there!!
Lets all do our share in "learning" .... all info is out there, waiting to be understood and learned....This is a forum where you can choose to build or not to...
SORRY, but im not showing/releasing my selection/parts optimized, w/c I have learned in a few months time... that would be unfair on my part (not unless I get paid?? :D) .. there is really NO OTHER option but to (build it) and learn...there is no KIT for it either... In those few months , I came close to understanding how a UCD works :) I might NOT be an expert/or an engineer, BUT I can 9and DID) learn...
BUILD the circuit from what chris has posted (and me).. I have posted what I have been using in my thread .... believe me, it WILL WORK!! just post your results and we could hopefully help you out...
-simulate- and build.... and learn... trust me, its FUN!! :D
Im curently making MY OWN -modulator- PCB... for a more nicer look... its DONE already but have not transfered it yet to board....
hope this clears up a bit...peace
Regards,
Raff |
|
|
| classd4sure |
Raff,
:)
I think part of the obvious difference is you went into it willing to learn /try /research, get your hands dirty and you were willing to accept even cryptic advice and made the effort to decipher it.
Can't have it all handed to you and if you did you wouldnt' have learnt. You'd be stuck with an amp that kind of sort of half works and wouldnt 'know what to do with it, then when you're fed up of swapping parts aimlessly without knowing what they're doing or why they're doing it, you'd be here picking bones, moaning about all the "experts" who dont' really know anything.
Can't please them all you know. Great idea to keep your final schematic personal, you worked hard enough for it, and anyway, people just use whatever parts they have around anyway, and then wonder why the values you've shown them don't work with their transistors.
There's only a dozen links and more schematics around here already that fully discuss the issue, long before the existance of the philips application note.
:rolleyes:
+- 55 hm..... I think I see a challenge :D |
|
|
| SSassen |
| quote: | | looks like you are an expert in experimenting cluelessly. |
I'd say you're plain wrong and have quite some nerve stating that, if there's someone that deserves to be padded on the back and complimented on all his work with a UcD style class-D amplifier then it is Chris. Esp. considering all people he's helped making their first endeavours into DIY class-D amplifiers, or understanding how they work in general.
| quote: | | you did a good job, only in the weakest possible sense. |
Quite the opposite to be honest, given the fact that all Chris had to work with is his simulator and a box of parts. He managed to get a UcD amplifier up and running, built from scratch using run of the mill components without the luxuries many of us take for granted such as a ocilloscope.
And what did you build exactly?
Best regards,
Sander Sassen
Positively not RoHS compatible |
|
|
| classd4sure |
Hey Sander,
Thanks.
I'm not all that bothered by him, feels like a gnat gnawing on my ankles is all.
You know what I still have that homebrew sucker in a semi working state... just needs a few caps that I borrowed for something.
Since I finally have a scope now, as soon as I make some probes and such..... I can't wait to throw that ol thing on the scope and see how frightening it is!!
I'll do all I can to get some pics of the gate waves from it and throw them up here so we may all laugh :D
BTW I never would have gotten as far as I did with it without Bruno's help! :worship: and many others like Charles Ivan Johan and I just like to give back a little.
Cheers |
|
|
| RX5 |
Chris,
post pic of your semi working.. :D
and your scope.... :smash:
anyways, off to make another board.. sort of daughhterboard for UCd type.. :) though not SMT, i wILL compress parts layout so i could make modulator board small....
:smash: |
|
|
| EC8010 |
| quote: | Originally posted by fokker
looks like you are an expert in experimenting cluelessly.
you did a good job, only in the weakest possible sense. |
:cop: :cop: :cop: :cop: :cop:
You are perfectly entitled to hold whatever opinion you like, and this forum exists to encourage constructive discussion of technical matters. If you wish to disagree with or dispute a technical point, then please do so without making it personal.
:cop: :cop: :cop: :cop: :cop: |
|
|
| ray_moth |
| EC8010 put it perfectly. I'm glad he got in before I did - I might have been a little less diplomatic. |
|
|
| fokker |
| quote: | Originally posted by SSassen
a box of parts |
that does make swapping out the parts that much easier, doesn't it? |
|
|
| rogs |
| quote: | Originally posted by rogs
The overcurrent protection circuit for this application is not included on the main circuit diagram on page 4, nor are the components included on the demo pcb. Probably not essential for simple testing, but useful to try.
The additional circuitry is shown on page 9 --- but I can't see how it works. It looks to me as if the zener diodes D1 and D2 are drawn in reverse - surely not the same polarity for both positive and negative halves?
Or am I missing something? |
Anyone ? |
|
|
| classd4sure |
| quote: | Originally posted by fokker
that does make swapping out the parts that much easier, doesn't it? |
talk to me after you've tried soldering new leads onto a used transistor dug out of an old component.
I feel you're in way over your head here, and perhaps are just too dense to notice. |
|
|
| EC8010 |
| quote: | Originally posted by classd4sure
and perhaps are just too dense to notice. |
:cop: :cop: :cop: :cop: :cop: :cop:
The "no personal attacks" rule applies to everybody. I'm basically a lazy person; please don't cause me extra work.
:cop: :cop: :cop: :cop: :cop: :cop: |
|
|
| classd4sure |
| Good job. I keep forgetting to make the coffee first and post after. :D |
|
|
| lumanauw |
Hi, Chris,
Is it possible to make this UCD amp a follower (G=1)? In the schematic, do not use R12/R11, output is directly attached to inverting input (like opamp buffer, to base of Q2).
What will happened? Will it blow out or not oscilating at all? |
|
|
| classd4sure |
Hi,
My guess would be some sort of chaotic mess if you tried it like that.
The proportional gain of the output stage is set by R12 and R11, so if you want it to be a gain of one, just use R12=R11. |
|
|
| RX5 |
add to that:
higher value R12 = higher GAIN
lower value R11 = higher GAIN
:smash: |
|
|
| lumanauw |
Hi, Chris,
R11 and R12 are voltage divider. If we use R11=R12, then gain G=2.
Is it impossible to make it a follower (G=1)? |
|
|
| djQUAN |
| why would you want it to be a follower anyway?:confused: |
|
|
| IVX |
| quote: | Originally posted by djQUAN
why would you want it to be a follower anyway?:confused: |
I guess, for more loop gain.:cool: |
|
|
| classd4sure |
| quote: | Originally posted by lumanauw
Hi, Chris,
R11 and R12 are voltage divider. If we use R11=R12, then gain G=2.
Is it impossible to make it a follower (G=1)? |
Right you are! Sorry, In my little world it would have worked that way :clown:
I always have the differential version in mind, and I recommend you use that instead.
You'll have 2XR11 and 2XR12, R11=R12, A~1 |
|
|
| Pafi |
| quote: | Originally posted by lumanauw
Hi, Chris,
Is it possible to make this UCD amp a follower (G=1)? In the schematic, do not use R12/R11, output is directly attached to inverting input (like opamp buffer, to base of Q2).
What will happened? Will it blow out or not oscilating at all? |
Oscillation freq will be much lower depending on load impedance, and Q1-Q2 may will be blown up because of high feedback voltage. |
|
|
| IVX |
| quote: | Originally posted by Pafi
Oscillation freq will be much lower depending on load impedance, and Q1-Q2 may will be blown up because of high feedback voltage. |
I don't believe that such apocalypse will occur yet. :cannotbe: |
|
|
| classd4sure |
| quote: | Originally posted by IVX
I don't believe that such apocalypse will occur yet. :cannotbe: |
I tried the single ended version with the feedback components removed it was pretty much a chaotic mess, I think Pafi's prediction is right on. First it really just slammed the output to the negative rail where it stuck, so I equalized the inverting/non inverting impedances and it oscillated very slowly, 50% duty cycle, output in full saturation (20V beyond the rails).
I haven't put much thought into it but there may be some chance at getting it to work if you configure it in non inverting mode, what do you think? Still, differential is vastly superior. |
|
|
| fokker |
| quote: | Originally posted by Pafi
and Q1-Q2 may will be blown up because of high feedback voltage. |
and that's something not many people appreciate in class d amp. |
|
|
| IVX |
| quote: | Originally posted by classd4sure
I tried the single ended version with the feedback components removed it was pretty much a chaotic mess, I think Pafi's prediction is right on. First it really just slammed the output to the negative rail where it stuck, so I equalized the inverting/non inverting impedances and it oscillated very slowly, 50% duty cycle, output in full saturation (20V beyond the rails).
I haven't put much thought into it but there may be some chance at getting it to work if you configure it in non inverting mode, what do you think? Still, differential is vastly superior. |
I'm talking about Fsw. Really, differential is vastly superior, but it's other matter, if comparator ready to work well (e.g. comparator VDD>VDD power) , so no problem. |
|
|
| Pafi |
IVX!
Fosc is the freq where loop phase shift reaches -180 degrees. This phase shift comes from 3 parts: filter, feedback, delay of comparator+powerstage. If we have feadback=1, then only 2 parts left. Delay of comp+PS introduces a small, negative component (phi=360*delay*freq). Worst case: an unloaded LC filter already has -180 phase shift above resonance freq, so with no load fosc=fresonance. I dont have to tell what does this mean. (Theoretically unlimited output voltage.)
With load it is less catastrophic, but still bad. You can simulate it with an ideal comparator, delay, and LC filter!
This is exactly why post-filter feedback is not trivial to design. |
|
|
| IVX |
| quote: | Originally posted by Pafi
IVX!
Fosc is the freq where loop phase shift reaches -180 degrees. This phase shift comes from 3 parts: filter, feedback, delay of comparator+powerstage. If we have feadback=1, then only 2 parts left. Delay of comp+PS introduces a small, negative component (phi=360*delay*freq). Worst case: an unloaded LC filter already has -180 phase shift above resonance freq, so with no load fosc=fresonance. I dont have to tell what does this mean. (Theoretically unlimited output voltage.)
With load it is less catastrophic, but still bad. You can simulate it with an ideal comparator, delay, and LC filter!
This is exactly why post-filter feedback is not trivial to design. |
Ok, let's make simulation, BTW, it's stable even completely without load.. AC analysis was done with stepping R1 from 8.2k to 200ohm. :cool: |
|
|
| classd4sure |
Hi Ivan,
Your simulation doesn't seem to match the discussed conditions. |
|
|
| IVX |
| quote: | Originally posted by classd4sure
Hi Ivan,
Your simulation doesn't seem to match the discussed conditions. |
I thought the moot point is G<=1 whatever detail. |
|
|
| classd4sure |
nah, it was no feedback resistor in the given inverting mode, that's what Pafi and I were talking about.
Wicked sim's though :) |
|
|
| Pafi |
IVX!
This was the question:
| quote: | | Is it possible to make this UCD amp a follower (G=1)? In the schematic, do not use R12/R11, output is directly attached to inverting input (like opamp buffer, to base of Q2). |
BTW: is UcD really so fast? (I mean is td=220ns real? I haven't built it yet.) |
|
|
| classd4sure |
| If I recall correctly and I probably don't, 200ns for the comparator, 250ns output stage. |
|
|
| IVX |
| quote: | Originally posted by classd4sure
If I recall correctly and I probably don't, 200ns for the comparator, 250ns output stage. |
So 200+250=450? Sure 200 or less even. AC show 330Khz vs 384Khz transient. |
|
|
| Pafi |
| Based on calculations 220 ns is believable, but is it real? Have anybody measure it? |
|
|
| IVX |
| quote: | Originally posted by Pafi
Based on calculations 220 ns is believable, but is it real? Have anybody measure it? |
Why need to measure, if is believable already? :) |
|
|
| fokker |
| quote: | Originally posted by Pafi
Have anybody measure it? |
how can anyone measure it if setting it to g=1 will blow the comparator transistors as suggested earlier? |
|
|
| classd4sure |
| quote: | Originally posted by Pafi
Based on calculations 220 ns is believable, but is it real? Have anybody measure it? |
I might do that once I get some probes.
Ivan, I looked around for Bruno's old post where he said what the delays were.... no luck in finding it. What software do you use for simulating? |
|
|
| IVX |
| quote: | Originally posted by classd4sure
I might do that once I get some probes.
Ivan, I looked around for Bruno's old post where he said what the delays were.... no luck in finding it. What software do you use for simulating? |
In the Bruno's patent 200nS too. Micro-Cap simulator, free demo version is available. |
|
|
| Pafi |
IVX!
| quote: | | Why need to measure, if is believable already? |
1.: Just because! :-)
2.: Delay spec. of IR2113 is believable too, even though I measured it much worse then spec. |
|
|
| IVX |
| quote: | Originally posted by Pafi
IVX!
1.: Just because! :-)
2.: Delay spec. of IR2113 is believable too, even though I measured it much worse then spec. |
I can say that my ir2110s (0443) gave /150ns&\112ns through 10ohm to 3500nF, it seems in the claimed tolerance. |
|
|
| blmn |
| quote: | Originally posted by Steven
Philips released its UCD application note.
http://www.semiconductors.philips.c...s/UM10155_1.pdf
The UcD concept was patented by Philips. However, the use of the IP is granted (without a license fee) to third parties that use the enabling semiconductor components from Philips Semiconductors in their application. |
Hi Steven,
I couldn´t find the file in the Philips site. Can you make it available here or send me a copy?
Regards |
|
|
| RX5 |
hmmm link got lost, how come? :xeye:
maybe another version to come.....:smash: |
|
|
| blmn |
HI RX5,
Your email provider returns account disable. Is it possible for you to attach the file on the forum?
Regards, |
|
|
| RX5 |
blmn,
link to that file is OK again :D
thanks to
Ouroboros :smash:
Regards,
Raff |
|
|
| blmn |
Tks all!!!
I will show you my first UcD, soon...
Best Regards, |
|
|
| lumanauw |
| If BF820/21, BC847/57 and all small transistors are not using SMD type, but using ordinary TO-92type, for example 300V rating using MPSA42/92, the 60V using MPSA06/56, for bigger ones using BD139/140, will this design still self-oscilating at 400-500khz, or to achieve this frequency it is only possible with SMD sized transistor? |
|
|
| RX5 |
| quote: | Originally posted by lumanauw
If BF820/21, BC847/57 and all small transistors are not using SMD type, but using ordinary TO-92type, for example 300V rating using MPSA42/92, the 60V using MPSA06/56, for bigger ones using BD139/140, will this design still self-oscilating at 400-500khz, or to achieve this frequency it is only possible with SMD sized transistor? |
thats possible...
:smash: |
|
|
| fokker |
this whole discussion of driver current output capabilities and large gate capacitance is interesting, particularly in comparison to those little transistors used by Phillips in that application note. If a 0.5ma or 3a gate driver current output isn't enough, how could those little transistors possibly do their job?
Some times we are so technical we don't see the forest. |
|
|
| lumanauw |
| But the small (SMD) transistor has small die area, making the Cob is relatively small following the size. TO-92 like MPSA42 has about 6pf, twice the SMD part. |
|
|
| RX5 |
| quote: | Originally posted by lumanauw
But the small (SMD) transistor has small die area, making the Cob is relatively small following the size. TO-92 like MPSA42 has about 6pf, twice the SMD part. |
as you can see in your UCD400, there is 1 MPSA92 there(daughterboard)... :) |
|
|
| zilog |
| I have a running UcD prototype using 2n5551, 2n5401, 2n3906, 2n3904 and BC640 etc. They are all non-optimal to-92:s, but still perform quite decently. Now I am trying to minimize my design and am thus moving to SMD for all parts that I can, but the bulkier design still works quite well. The only real change compared to the old design will be improved gate drivers (faster turn-off, same speed turn-on), the rest will only be about reducing PCB and high current loop area. |
|
|
| RX5 |
| quote: | Originally posted by zilog
I have a running UcD prototype using 2n5551, 2n5401, 2n3906, 2n3904 and BC640 etc. They are all non-optimal to-92:s, but still perform quite decently. Now I am trying to minimize my design and am thus moving to SMD for all parts that I can, but the bulkier design still works quite well. The only real change compared to the old design will be improved gate drivers (faster turn-off, same speed turn-on), the rest will only be about reducing PCB and high current loop area. |
thats cool... does your UCD proto include op-amp buffer? or just the modulator/mosfet drivers??
I assume u used 3906/3904 in place of 2222A and 2907?? :)
care to share some photos? |
|
|
|
