CurtPalme.com Home Theater Forum

koldby · Posted: Mon Jan 19, 2015 10:28 am Post subject: G70 video bandwidth

I have been looking at the G70 schematic to see if there was a way to increase the bandwith in the video chain.

As far as I can see the bottelneck is in the neck board Laughing

Isen´t the limiting factor the use of the G! drive amp VPH06 as this is rated to max 65 Mhz?

Actually I find it very strange, that Sony has chosen this amp for the G1 drive and the VPA15 for the cathode drive.
VPA 15 is rated to max 150 Mhz, but can only put out 40 V p-p while VPH06 can put out 100 V p-p.
But why has Sony found it nessesary to have a 100 V p-p amp for G 1?

Even the amps in Marquee 9500 only have - 85 V PS for the G1 drive (and + 85 V for the cathode).

Has anybody tried to use a VPA 15 for G1 (of course with reduced PS and some changes in the gain in the input amp)?

Or maybe a discrete solution as the Marquee or this :

http://www.datasheetlib.com/datasheet/849009/an1020_on-semiconductor.html

Excuse me if all this is just stupid questions, but I am here to learn... Embarassed

Koldby

barclay66 · Posted: Mon Jan 19, 2015 11:02 am Post subject:

Hi,

There are no stupid questions!
Franly, I don't know why Sony chose to implement a differential amplifier design and didn't use the same part for both ends. Electrohome instead went for a strictly symmetrical design which allowed for a good performance and reduction of peaking to the minimum.

Therefore, in my opinion, the best approach is with discrete transistors because:

- It allows for a wider range of suitable parts (Preamp an Power Transistors)
- It allows for tuning the circuit where needed
- It allows for taking measurements at each stage

Besides that, Sanyo has stopped production of the VPA/VPH/VPJ/VPM/VPS circuits more than 10 years ago. So it will be difficult to get the desired ones although some of them have interesting specs (see VPA25 in the datasheet).

Too bad that high-voltage/high-speed applications have become almost obsolete since the end of the CRT era. Nowadays it becomes more and more difficult finding transistors that have a high transition frequency (>800 MHz) and are rated for high voltages (>120 V) at the same time...

Regards,
barclay66

koldby · Posted: Mon Jan 19, 2015 11:43 am Post subject:

Thanks Barclay

I can see why a discrete solution is ultimately better, but VPA15 (or 25) would be a drop in replacement only requring a lowering of the supply voltage and some gain adjustment.
Would it be possible?
I haven´t quite found out what the circuit aftre the DC blocking capacitor C110 between VPH06 and G1 is doing. Looks like part of it reduces and stabilize the -180 V to a mor sane -G1 voltage? Seams there is some sort of max. Cathode current protection going on there too?
Could this be the reason Sony needs such a high swing?
Question

koldby

koldby · Posted: Mon Jan 19, 2015 11:43 am Post subject:

Thanks Barclay

I can see why a discrete solution is ultimately better, but VPA15 (or 25) would be a drop in replacement only requring a lowering of the supply voltage and some gain adjustment.
Would it be possible?
I haven´t quite found out what the circuit aftre the DC blocking capacitor C110 between VPH06 and G1 is doing. Looks like part of it reduces and stabilize the -180 V to a mor sane -G1 voltage? Seams there is some sort of max. Cathode current protection going on there too?
Could this be the reason Sony needs such a high swing?
Question

koldby

koldby · Posted: Mon Jan 19, 2015 11:44 am Post subject:

Thanks Barclay

I can see why a discrete solution is ultimately better, but VPA15 (or 25) would be a drop in replacement only requring a lowering of the supply voltage and some gain adjustment.
Would it be possible?
I haven´t quite found out what the circuit aftre the DC blocking capacitor C110 between VPH06 and G1 is doing. Looks like part of it reduces and stabilize the -180 V to a mor sane -G1 voltage? Seams there is some sort of max. Cathode current protection going on there too?
Could this be the reason Sony needs such a high swing?
Question

koldby

barclay66 · Posted: Mon Jan 19, 2015 12:15 pm Post subject:

Hi,

I think that a replacement should be possible. There's one forum member which even dared to modify the video packs themselves (he replaced the internal transistors with higher bandwith ones): https://www.curtpalme.com/forum_archived/viewtopic.php@t=34710.html
Unfortunately it's difficult to read what his translator produced.
As I don't have the G70 neck board schematics I can't help on your second question though...

Regards,
barclay66

mp20748 · Posted: Mon Jan 19, 2015 12:40 pm Post subject:

I've also looked at the G70 video chain, but would say the real problem would not be the neck boards. It's the stages before the neck boards is where the real problem is for bandwidth improvement.

The Sanyo modules are very difficult to work with. Some are interchangeable, and they are the ones that's close in bandwidth ratings. Anything of any real significance don't seem to work plug-and-play.

The downside of the Sanyo module is that each one or number require a specific external to it peaking circuit. This is where you could also run into problems, because you'd have to pull out a scope and spend some time tuning and tweaking.

So I would leave the neck boards alone, and look at the other part of the video chain, where you can really make some quick improvements.

It's been years since I've seen the diagram, but if the same people also designed the G90, you'll need to remove some really stupid stages from the video chain for the absolute shortest video chain. And do remember that some of the chips can stand an improvement or be replaced with higher bandwidth chips.

Can you post a picture of the video chain here?

barclay66 · Posted: Mon Jan 19, 2015 12:45 pm Post subject:

Nashou66 · Joined: 12 Jan 2007 Posts: 16171 Location: West Seneca NY

Cascading amps increases bandwidth. This is how the Marquee VNB's were done, but used RF transistors which is an amazing idea and a wonder of engineering. Those guys were very proud of that design.

The VIM I believe also cascades op amps. the older 03p vims had two RGB op amps while the newer 02p used another op amp built into the multiplier(AD835) the old multiplier AD834 did not have an op amp built in.

Here is an interesting video on that subject.

https://www.youtube.com/watch?v=ZvT9hHG17tQ

Maybe adding another on to a daughter board might help but I bet a lot of re engineering of the rest of the circuit would be needed. I bet Mike could do it ! Very Happy

Athanasios

koldby · Posted: Mon Jan 19, 2015 1:55 pm Post subject:

koldby · Posted: Mon Jan 19, 2015 2:02 pm Post subject:

barclay66 · Posted: Mon Jan 19, 2015 3:21 pm Post subject:

Hi,

From what I can see there are indeed some bottlenecks in the G70 video chain:

- The LM1283 has 140MHz bandwith only
- The EL2160 has 130/180MHz bandwith only depending on gain settings

This is quite low for a preamp stage. Normally You would expect something in the 500MHz to over 1GHz area...

Regards,
barclay66

mp20748 · Posted: Mon Jan 19, 2015 4:12 pm Post subject:

LM1283 if my memory is correct, is the same one in the G90. You can get a tad more bandwidth out of it depending on how they couple the outputs of it. The other chip is a must change device. It's where you should see your best increase.

koldby · Posted: Mon Jan 19, 2015 5:34 pm Post subject:

Thanks all.
I will start with finding some better video opamps and see if that helps-

gjaky · Joined: 05 Jun 2010 Posts: 2802 Location: Budapest, Hungary

The VPH06 might seem to be a 65Mhz chip but it has similar slew rate to the VPA15H, also note the H designation after the 15, that means it is a high-swing version, and also can stand higher supply voltage (something like 115-120V, while plain VPA15 only stands 90V).
The VPH06 is not used in the way how it was originally meant, but I find this a rather clever idea.

These Sanyo videopacks are containing a so called cascode circuit, these circuits are consuming less power if the input is 'black', and the consumption raises along with the brightness up to white, so far so good. With differentially driving both G1 and K on the tube one can gain speed but the driving signals have to be inverted for the two electrodes.
This means even if the tube output is 'black' one videopack works as black, but the other is working inverted ie. consuming power as if the output would be white. When the output turns to 'white' one ic works as according to it -> 'white', the other works as if it would do 'black'. This means either chip is always under high (heat) stress and this is the biggest problem with the symmetrical VPA configuration. I think this is why the VPH06 is there (that is the inverted ic) it is under less demanding conditions that way. For example in the NEC XG both ics are the same (VPJ13S) but their heatsink is always very hot, and they indeed fail time to time.
On the other hand tha marquee neckboard really has great construction, since it is fully symmetrical, so both sides only stressed at the same time.

gjaky · Joined: 05 Jun 2010 Posts: 2802 Location: Budapest, Hungary

Check this out as well: L I N K
Unfortunately some of the references are seem to be dead, but the point is still clear.

macgyver655 · Joined: 22 Aug 2007 Posts: 8508

gjaky · Joined: 05 Jun 2010 Posts: 2802 Location: Budapest, Hungary

koldby · Posted: Tue Jan 20, 2015 2:29 pm Post subject:

koldby · Posted: Tue Jan 20, 2015 3:42 pm Post subject:

One more thing.
I have not been able to find current feedback op-amps with 300+ Mhz BW AND +/- 12 V ps..? (substitute for el2160)