| Author |
Message |
Boilermaker
Joined: 21 May 2006
Posts: 527
|
| Posted: Fri Jan 20, 2012 8:49 pm Post subject: |
|
|
I know - I don't have a Marquee, and I only follow this thread because its fun BUT!!!
Don't forget the voltage ratings if you are removing caps from a series configuration - They are also frequency dependent.
Also, you are reducing ESR when you remove any from a series.
OK, now I'll go back to my NEC's.
Bob
|
|
| Back to top |
|
 |
draganm
Joined: 08 Mar 2006
Posts: 8990
Location: Colorado
|
| Posted: Fri Jan 20, 2012 9:44 pm Post subject: |
|
|
| Nashou66 wrote: | | Caps are a funny thing, if you look at the schematic there are caps in Parallel as in C4 and C5 but you also have caps in series and when in series capacitance decreases. So removing those two caps you take two caps out of series which will raise the total capacitance. |
ok, that's what i missed. They all looked like they're parallel to me. The question is did Dennis run the smaller caps? I don't think he could have because presumably his board is still running. Something here has been missed I think. Either the total capacitance is wrong or like boiler pointed out we have an ESR issue or? I wish Scott or Dennis would comment here 
I might just have to skip the Cree parts and concentrate on the bigger heat-sink.
| Boilermaker wrote: | | I know - I don't have a Marquee, and I only follow this thread because its fun BUT!!! |
yeah fun for you, you didn't just spend $150. on a board that now produces and error light 
| Boilermaker wrote: | | Don't forget the voltage ratings if you are removing caps from a series configuration - They are also frequency dependent. |
I kept the voltage rating at 1KV which is what came out. I don't think the caps I put in have failed completely. The machine begins a normal start-up but then shuts down the HV after a second. I bet If I could get it back to long retrace time it would run.
|
|
| Back to top |
|
|
Nashou66
Joined: 12 Jan 2007
Posts: 16171
Location: West Seneca NY
|
|
| Back to top |
|
|
draganm
Joined: 08 Mar 2006
Posts: 8990
Location: Colorado
|
| Posted: Fri Jan 20, 2012 10:13 pm Post subject: |
|
|
OK, I think I see it now. C4, 6, and 10 as well as 5,7, and 11 are in series. So my question is should we have left Q3 in, removed Q7, and left the bull bank of 6 caps alone?
| Description: |
|
| Filesize: |
117.07 KB |
| Viewed: |
7514 Time(s) |
|
|
|
| Back to top |
|
|
Boilermaker
Joined: 21 May 2006
Posts: 527
|
| Posted: Fri Jan 20, 2012 11:14 pm Post subject: |
|
|
| Quote: | | yeah fun for you, you didn't just spend $150. on a board that now produces and error light |
Yeah - It's always more fun spending someone else's money! LOL
| Quote: | | I kept the voltage rating at 1KV which is what came out |
.
Usually the only reason for putting caps in series is to increase the voltage capability, but you will end up with a reduction in capacitance value as Nash stated. I don't have the slightest idea what you are doing, but for example:
Assuming all caps are of equal capacitance, if you have three caps in series across 1,000 volts, each cap will have 333 volts across it. If you remove one of the caps to effectively increase the total capacitance, each cap will now have 500 volts across it. With just one cap, it will be 1,000 volts. Just wanted to warn you and let you know that this assumes all caps have the same value.
Bob
|
|
| Back to top |
|
|
draganm
Joined: 08 Mar 2006
Posts: 8990
Location: Colorado
|
| Posted: Sat Jan 21, 2012 1:26 am Post subject: |
|
|
| Boilermaker wrote: | Usually the only reason for putting caps in series is to increase the voltage capability, but you will end up with a reduction in capacitance value as Nash stated. I don't have the slightest idea what you are doing, but for example:
Assuming all caps are of equal capacitance, if you have three caps in series across 1,000 volts, each cap will have 333 volts across it. If you remove one of the caps to effectively increase the total capacitance, each cap will now have 500 volts across it. With just one cap, it will be 1,000 volts. Just wanted to warn you and let you know that this assumes all caps have the same value.
Bob |
well dammit Bob go back and read it all starting at page 38 or so, I need help here 
That makes sense, they did put 3 x 1KV rated caps in series from what I can see. The pulses on the H-sweep are in the 800Volt range AFAIK but I'm guessing transients can be double that or more. According to Scott
| Quote: | | The original HDM has three 600V mosfets connected in series for a blocking capability of 1800V. Two of the 1200V Cree parts could block 2400V which is an improvement over the original circuit. Three of these in series is not necessary. Dissipation could be lowered with two.There are other parts on the HS contributing to the heat. The width regulator is a series pass type circuit kinda like a LM317. It regulates by dropping voltage across the resistance of Q9 and especially Q10. |
So it's possible the circuit might see spikes in the 2KV range? In which case a pair of 1KV caps in series as opposed ot 3 would be bad news. Dammit, if Scott or Dennis have nothing to add to this it's dead in the water, bummer.
|
|
| Back to top |
|
|
Nashou66
Joined: 12 Jan 2007
Posts: 16171
Location: West Seneca NY
|
|
| Back to top |
|
|
Nashou66
Joined: 12 Jan 2007
Posts: 16171
Location: West Seneca NY
|
|
| Back to top |
|
|
tse
Joined: 03 May 2006
Posts: 1014
Location: Sweatbucket, Fl.
|
| Posted: Sat Jan 21, 2012 2:24 am Post subject: |
|
|
Draganm,
You should be working with something like this. Put the cap values that you are using in this drawing. You could try starting the projector without an input so it defaults to 31.5kHz with the deflection yoke sections in series to see if it works with that condition.
Scott
| Description: |
|
| Filesize: |
5.48 KB |
| Viewed: |
7469 Time(s) |
|
_________________
"Were we directed from Washington when to sow and when to reap, we would soon want bread."
Thomas Jefferson
|
|
| Back to top |
|
|
draganm
Joined: 08 Mar 2006
Posts: 8990
Location: Colorado
|
| Posted: Sat Jan 21, 2012 6:59 pm Post subject: |
|
|
| Nashou66 wrote: | Drags look back and find where Scott talked about the voltage needed to do the sweeps( He had a spice mold comparing Dennis's cap values), it was the three fets added together which was I think 1800Vdc, the Crees are what 800 Each? I forgot but two Crees voltage added is the top voltage those Caps might see.
Athanasios |
Nash the big FET's are specced for voltage blocking (Speed reading will not help here)  , I quoted the section about it just above your reply. Based on the original configuration 1800 volts was required, 2400 volts is the new capability/capacity of the circuit.
Scott thanks for the clean drawing, this configuration worked but I had those weird angled tails on the raster /test pattern.
|
|
| Back to top |
|
|
draganm
Joined: 08 Mar 2006
Posts: 8990
Location: Colorado
|
| Posted: Sat Jan 21, 2012 7:01 pm Post subject: |
|
|
|
this caused H-fail
|
|
| Back to top |
|
|
Nashou66
Joined: 12 Jan 2007
Posts: 16171
Location: West Seneca NY
|
|
| Back to top |
|
|
draganm
Joined: 08 Mar 2006
Posts: 8990
Location: Colorado
|
| Posted: Sat Jan 21, 2012 7:31 pm Post subject: |
|
|
| Nashou66 wrote: | Drags Try using all 3.9nf caps.
then maybe add a little more in parallel till you find the best value for the known resolution you plan on using.
You might have to compromise a specific value if you plan to use multiple res's.
Athanasios |
that's not a bad idea, but I'll wait for Scott to reply. I don't really need multiple resolution capability on this, something that will work specifically with 1080P60 or 72Hz and short retrace would be just fine.
|
|
| Back to top |
|
|
tse
Joined: 03 May 2006
Posts: 1014
Location: Sweatbucket, Fl.
|
| Posted: Sat Jan 21, 2012 8:33 pm Post subject: |
|
|
The caps in the works image should be equivalent ot about 5.35nF In the hfail one about 3.05nF.
I plugged them into my Mathcad worksheet (not exact but pretty close) and nothing looks dangerous.
The flyback caps dump their charge into the yoke to produce the first half of the deflection line. Maybe 3.05nF isn't enough cap?
It shouldn't hurt anything to put 3.9nF caps in place of the 2.2nF ones presently installed.
Scott
| Description: |
|
| Filesize: |
7.79 KB |
| Viewed: |
7385 Time(s) |
|
| Description: |
|
| Filesize: |
7.77 KB |
| Viewed: |
7389 Time(s) |
|
_________________
"Were we directed from Washington when to sow and when to reap, we would soon want bread."
Thomas Jefferson
|
|
| Back to top |
|
|
draganm
Joined: 08 Mar 2006
Posts: 8990
Location: Colorado
|
| Posted: Sun Jan 22, 2012 1:53 am Post subject: |
|
|
| tse wrote: | | It shouldn't hurt anything to put 3.9nF caps in place of the 2.2nF ones presently installed.Scott |
thanks again Scott. I can actually pull those off some parts boards here and try them without waiting for another mouser delivery
| tse wrote: | | Maybe 3.05nF isn't enough cap? |
it's interesting that the schematic dennis posted, which is presumably a late model 2003-08P board , had a pair of 4n7's and only one 2n2 in the secondary location. Since they might have been trying to lower retrace time as much possible, maybe this was the smallest value they could use while maintaining proper function.
|
|
| Back to top |
|
|
dvh99
Joined: 25 Dec 2009
Posts: 2158
Location: nederland
|
| Posted: Tue Jan 24, 2012 3:04 pm Post subject: |
|
|
it`s a trade of as you can see in the calculations.
i think the newer switching mosfets (apt parts iirc) could handle higher voltages so shorter retrace timings could be used, this meant going with a lower capacitance for the flyback cap.
the cree can handle up to 2400v with 2 in series so they are safe.
edit: the one thing that annoys me however is the flimsy connection of the hdm module to the motherboard.
my image was shaking horizontally and of course a cleaning and reinstalling solved this.
with a board that carries such high voltages and current peaks one would expect a much firmer and sturdier connection.
_________________
1 answer always poses multiple questions.
marquee 9500ultra HD10L moome hdmi1.3 v3+ some mods.
|
|
| Back to top |
|
|
draganm
Joined: 08 Mar 2006
Posts: 8990
Location: Colorado
|
| Posted: Tue Jan 24, 2012 5:36 pm Post subject: |
|
|
well not sure what's going on with this dam thing, putting in larger poly caps didn't fix it. the sets starts up normally but as soon as the HV energizes it shuts back off and fades away, then an H-fail light.
Something died the second time I installed this board, there was noise on the tube face just before it happened too.
My guess is either one of the Cree fet's or Clare drivers.
Dennis can you please tell me
1) what value caps are you running exactly in the fly-back positions
2) are you running long or short retrace
3) if short, do you see the "tails" on the test pattern like this (roughly to scale at 720P, smaller at 1080P)
|
|
| Back to top |
|
|
dvh99
Joined: 25 Dec 2009
Posts: 2158
Location: nederland
|
| Posted: Tue Jan 24, 2012 6:31 pm Post subject: |
|
|
1: 2n2 and 3n9 at both fets now (replaced the 4n7).
2: short retrace.
3: not a single tail to be seen, in fact it`s pretty damn straight  .
i have these fets installed on a later model ultra hdm of course.
_________________
1 answer always poses multiple questions.
marquee 9500ultra HD10L moome hdmi1.3 v3+ some mods.
|
|
| Back to top |
|
|
draganm
Joined: 08 Mar 2006
Posts: 8990
Location: Colorado
|
| Posted: Tue Jan 24, 2012 9:45 pm Post subject: |
|
|
| dvh99 wrote: | | i have these fets installed on a later model ultra hdm of course. |
what difference would that make?
at any rate, it's good that it worked out for you.
|
|
| Back to top |
|
|
dvh99
Joined: 25 Dec 2009
Posts: 2158
Location: nederland
|
| Posted: Tue Jan 24, 2012 9:54 pm Post subject: |
|
|
only other thing that is different is that i have the uc3709 drivers.
you did replace both u1 and u12 and lifted the wire from the yoke connector out of the pcb where the 24v goes to?
_________________
1 answer always poses multiple questions.
marquee 9500ultra HD10L moome hdmi1.3 v3+ some mods.
|
|
| Back to top |
|
|
|
|
