EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: james_s on February 20, 2019, 10:24:39 pm
-
I've reverse engineered the horizontal deflection and HV circuit of a common Chinese B&W portable CRT TV set. The end goal is to come up with a robust and reliable HV board for my 5" vector monitor project that uses the CRT from one of these sets but in the process I'd also like to fully understand how the existing circuit works and expand my knowledge in the area of SMPS design.
Anyway looking at the existing circuit, am I correct in thinking this is a L-C-L resonant tank with the flyback primary, C1, and the horizontal deflection yoke? What is the purpose of the arrangement with D1 and C9, and also C8, C10 and D2?
Now I wired up the circuit outside of the TV, using a mosfet in place of the original HOT and it works fine. Actually being built on a copperclad ground plane the waveforms look significantly cleaner than they do in the TV. What I'd like to do though is achieve the same result without needing the large inductance provided by the yoke, better yet if I could get away with no separate inductor at all. Driving the flyback transformer alone I was able to get about 5kV on the anode using a 12V input but that was pushing the duty cycle up quite close to where it seems the core would saturate and input current shoots way up. Any experts care to weigh in and nudge me in the right direction? I've done plenty of HV experiments driving flybacks but normally I'm trying to get high output, here I'm more interested in stable output, long term reliability and reasonable efficiency. Also if at all possible I would like to utilize only the original windings in the flyback instead of adding any new ones to the core.
-
D1 is commonly referred to as "boost diode". It allows current to flow into pin 1 only during the scan period. The tap at pin 1 causes the transformer primary to act as an auto-transformer, boosting the voltage at pin 2 to higher than supply voltage. It's this boosted voltage that the yoke inductance sees during scan to achieve the required deflection current.
The over-wind between pins 8 and 10 keeps the voltage at pin 8 at about 0V while D2 is conducting during the period between the start of the scan and when the transistor gets turned on. This is to prevent a step change in yoke voltage which would cause a noticeable change in scan rate.
-
Interesting thanks, I suspected that diode and capacitor may act like a boost converter but I wasn't certain whether there was more to it. There is about 15V at that pin in the original TV which supports that being the case. I had no idea what D2 was for, that's interesting. I've worked on a lot of CRT displays but it was never necessary to understand the horizontal circuit in depth to repair them.
-
I read somewhere that a technique is used to improve EHT load regulation by designing the EHT winding to be self resonant to some particular multiple of the flyback tuning frequency. I suspect this is also the reason that I have often seen a double hump at the top of the flyback pulse.
It may be best to keep the flyback tuned frequency and primary impedance unaltered. The scan rate isn't critical to this, it will only affect the flyback and EHT voltage.
-
That's interesting, indeed there is a double hump when used in the original TV.
I spent some more time playing with this tonight. I suspect that it will have to be a resonant design in order to get the desired output voltage. I tried driving the flyback directly over a broad range of frequency and duty cycle and I was never able to get more than about 5kV out of it even up to the point where the core is starting to saturate. I removed C9 and that seemed to have little effect on the this application, and I also experimented with D2/C8/C10. Curiously the diode also makes little difference in the HV, but with nothing connected to that pin of the flyback transformer the waveforms become a complete mess. Connecting even just one of those little film capacitors between pin 10 and ground cleans it up tremendously, I don't yet fully understand why.
Digging through my inductor stash I found one that when placed in series with the original C1 lets me get 7kV easily at only about 9V input and a drive frequency of about 28kHz which is well out of the audible range so even if I don't come up with something better, that's a possible solution.
-
For best results, the yoke inductance needs to be substituted with a gapped ferrite inductor that can be easily adjusted. The idea is to keep the flyback tuned frequency the same as before.
The flyback pulse is actually a bit over a half of a cycle of resonance between the lumped primary side inductance and the tuning capacitors, C8 and C10.
C1 can be omitted if the inductor is wired directly across the primary winding. During flyback, a large current circulates between the inductor and the tuning capacitors. This current path will include a bypass capacitor that will be needed between pin 2 and ground.
There is no real need for D2 if a MOSFET is used because of the body diode. There might be ringing at pin 10.
The drive duty cycle isn't critical. The MOSFET can be turned on at any time from after the flyback pulse drops to 0V to before the inductor current reverses, at somewhere just before half scan. Turning it on earlier will improve efficiency slightly. Avoid turning it on during the flyback pulse.
There is no need for D1 if a higher voltage is applied to pin 2.
-
If the flyback core is saturating, and there's no net DC current, try increasing the frequency. If there is net DC current and the core is ungapped, then the core flux needs to be reset on a cycle by cycle basis. I expect that the primary current gets shuffled off to some reactive network somewhere and then comes back in for the next cycle. I remember seeing all sorts of extra stuff in the H drivers of the old CRT multisync monitors to get them to work over a wide range. As to design information, I sort of recall seeing some horizontal sweep design notes in some of the old RCA semiconductor data and application books.
Is the secondary loaded correctly? If not, the saturation could be a voltage regulation feature. Remember that the flux in the core is a function of the primary amp-turns putting in less the seconday amp-turns taking out so that the load is an integral part of operation.
Best o' luck,