Forums

LTC6992 TimerBlox, jitter at low duty cycles

Started by Joerg January 29, 2015
Folks,

Got an LTC6992 in the -3 version which goes from 0% to 95% duty cycle, 
controlling a fairly massive PWM stage. It lives inside a (stable) PID 
loop and works ok except at very low duty cycles under 2%. There is 
breaks into a jitter and because I am driving big capacitive loads that 
results in tons of wasted power. Also, it quite abruptly turns off at a 
fraction of a percent instead of gracefully going towards 0%.

I've babied the device quite well and all the signals look clean, 
including on the very sensitive SET pin. In case stuff gets into the die 
I have tried shielding the whole stage and even that changed nothing.

Anyone had that happen and knows a fix?

-- 
Regards, Joerg

http://www.analogconsultants.com/
On Thu, 29 Jan 2015 15:52:43 -0800, Joerg <news@analogconsultants.com>
wrote:

>Folks, > >Got an LTC6992 in the -3 version which goes from 0% to 95% duty cycle, >controlling a fairly massive PWM stage. It lives inside a (stable) PID >loop and works ok except at very low duty cycles under 2%. There is >breaks into a jitter and because I am driving big capacitive loads that >results in tons of wasted power. Also, it quite abruptly turns off at a >fraction of a percent instead of gracefully going towards 0%. > >I've babied the device quite well and all the signals look clean, >including on the very sensitive SET pin. In case stuff gets into the die >I have tried shielding the whole stage and even that changed nothing. > >Anyone had that happen and knows a fix?
Usually noise in the Vcc supply, or inherient in the device operation. Does it do it on a seperate proto board? I had a IRF 1/2 cycle PFC part that was noisy, tuff to decouple. A compatible TI part was well behaved. Cheers
Den fredag den 30. januar 2015 kl. 00.52.08 UTC+1 skrev Joerg:
> Folks, > > Got an LTC6992 in the -3 version which goes from 0% to 95% duty cycle, > controlling a fairly massive PWM stage. It lives inside a (stable) PID > loop and works ok except at very low duty cycles under 2%. There is > breaks into a jitter and because I am driving big capacitive loads that > results in tons of wasted power. Also, it quite abruptly turns off at a > fraction of a percent instead of gracefully going towards 0%. > > I've babied the device quite well and all the signals look clean, > including on the very sensitive SET pin. In case stuff gets into the die > I have tried shielding the whole stage and even that changed nothing. > > Anyone had that happen and knows a fix? >
some RC+diode concoction to "subtract" some duty cycle? so it'll do 0-90% cleanly? -Lasse
On 2015-01-29 4:17 PM, Martin Riddle wrote:
> On Thu, 29 Jan 2015 15:52:43 -0800, Joerg <news@analogconsultants.com> > wrote: > >> Folks, >> >> Got an LTC6992 in the -3 version which goes from 0% to 95% duty cycle, >> controlling a fairly massive PWM stage. It lives inside a (stable) PID >> loop and works ok except at very low duty cycles under 2%. There is >> breaks into a jitter and because I am driving big capacitive loads that >> results in tons of wasted power. Also, it quite abruptly turns off at a >> fraction of a percent instead of gracefully going towards 0%. >> >> I've babied the device quite well and all the signals look clean, >> including on the very sensitive SET pin. In case stuff gets into the die >> I have tried shielding the whole stage and even that changed nothing. >> >> Anyone had that happen and knows a fix? > > Usually noise in the Vcc supply, or inherient in the device operation. > Does it do it on a seperate proto board? >
It did. Tried battery operation as well. VCC is very clean.
> I had a IRF 1/2 cycle PFC part that was noisy, tuff to decouple. A > compatible TI part was well behaved. >
I don't think anyone make a compatible part to the LTC6992. -- Regards, Joerg http://www.analogconsultants.com/
On 2015-01-29 4:23 PM, Lasse Langwadt Christensen wrote:
> Den fredag den 30. januar 2015 kl. 00.52.08 UTC+1 skrev Joerg: >> Folks, >> >> Got an LTC6992 in the -3 version which goes from 0% to 95% duty cycle, >> controlling a fairly massive PWM stage. It lives inside a (stable) PID >> loop and works ok except at very low duty cycles under 2%. There is >> breaks into a jitter and because I am driving big capacitive loads that >> results in tons of wasted power. Also, it quite abruptly turns off at a >> fraction of a percent instead of gracefully going towards 0%. >> >> I've babied the device quite well and all the signals look clean, >> including on the very sensitive SET pin. In case stuff gets into the die >> I have tried shielding the whole stage and even that changed nothing. >> >> Anyone had that happen and knows a fix? >> > > some RC+diode concoction to "subtract" some duty cycle? so it'll do 0-90% cleanly? >
92% would suffice. I thought about that but tossed the idea because it would be a major hack and re-layout. Then I might as well build my own PWM modulator. -- Regards, Joerg http://www.analogconsultants.com/
On Thu, 29 Jan 2015 15:52:43 -0800, Joerg <news@analogconsultants.com>
wrote:

>Folks, > >Got an LTC6992 in the -3 version which goes from 0% to 95% duty cycle, >controlling a fairly massive PWM stage. It lives inside a (stable) PID >loop and works ok except at very low duty cycles under 2%. There is >breaks into a jitter and because I am driving big capacitive loads that >results in tons of wasted power. Also, it quite abruptly turns off at a >fraction of a percent instead of gracefully going towards 0%.
That's on the data sheet, bottom of p 9. Some, not all, parts crash to 0% duty cycle at about 0.1 volt in.
> >I've babied the device quite well and all the signals look clean, >including on the very sensitive SET pin. In case stuff gets into the die >I have tried shielding the whole stage and even that changed nothing. > >Anyone had that happen and knows a fix?
The specs on that part are awful. Don't know how they managed to make it that bad. -- John Larkin Highland Technology, Inc picosecond timing laser drivers and controllers jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
On 2015-01-29 7:56 PM, John Larkin wrote:
> On Thu, 29 Jan 2015 15:52:43 -0800, Joerg <news@analogconsultants.com> > wrote: > >> Folks, >> >> Got an LTC6992 in the -3 version which goes from 0% to 95% duty cycle, >> controlling a fairly massive PWM stage. It lives inside a (stable) PID >> loop and works ok except at very low duty cycles under 2%. There is >> breaks into a jitter and because I am driving big capacitive loads that >> results in tons of wasted power. Also, it quite abruptly turns off at a >> fraction of a percent instead of gracefully going towards 0%. > > That's on the data sheet, bottom of p 9. Some, not all, parts crash to > 0% duty cycle at about 0.1 volt in. >
Yes, but it doesn't say that it exhibits a tarantella dance around that point.
>> >> I've babied the device quite well and all the signals look clean, >> including on the very sensitive SET pin. In case stuff gets into the die >> I have tried shielding the whole stage and even that changed nothing. >> >> Anyone had that happen and knows a fix? > > The specs on that part are awful. Don't know how they managed to make > it that bad. >
I really like LTC as a supplier but that part sure has me non-plussed. On the production run it's going to go bye-bye. -- Regards, Joerg http://www.analogconsultants.com/