6128 Plus video output distribution amplifier (splitting)

[Cwiiis]

I'm posting in the hope that someone has successfully done this, or knows how to go about it. I want to stream development from my Plus and still use the original monitor. I happen to have a Plus -> SCART adapter and a SCART -> CM14 adapter and going straight through works fine, as you might expect (the cables are passive, so this seems like an obvious outcome). Unfortunately, the signal coming out of the Amstrad is really far out of spec it seems... I got a SCART distribution amplifier from JS Technology (https://www.js-technology.com/store/index.php?id_product=59&controller=product) but the Amstrad monitor doesn't like the signal coming out of it. Blues are much darker, white screens go grey and bright screens start losing sync. I ran through several steps debugging it with them and at the end of it, though I've made some mods to the amplifier that make the situation a little better (sync is better, white is white), it's still not good enough (blue is still almost black, sync still starts skewing on bright screens).

So, has anyone managed to successfully duplicate video output from a Plus? Ideally, I'd want one output to go to a CM14 and then another output (really of any kind, analog or digital) that I can use with my capture equipment. If anyone knows what kind of modification or circuit would allow me to do this, I'd be extremely grateful!

[Bryce]

There are quite a few schematics online for active VGA splitters. Most of the circuits will work fine to duplicate the CPC output for 2 or more monitors.

Bryce.

[Cwiiis]

So, ignoring the ID pins and figuring something out for audio, I suppose this would work? It seemed that the impedance of the RGB channels was 100 ohms on the Amstrad rather than 75, would the resistance values need to differ?

[RetroCPC]

Yes correct, the RGB needs to be terminated by 100 Ohms for correct Colour "Mid Level Colour Intensity".

Attached is a modified Schematic, with two styles of RGB buffers - (one using diodes and the other using transistors connected as diodes) - the later offering better buffered Voltage offset accuracy.

Also, a slightly differant circuit for the Video Sync buffer circuit..


I've not tested the circuit but should be ok... (hopefully).

[Cwiiis]

Yes correct, the RGB needs to be terminated by 100 Ohms for correct Colour "Mid Level Colour Intensity".

Attached is a modified Schematic, with two styles of RGB buffers - (one using diodes and the other using transistors connected as diodes) - the later offering better buffered Voltage offset accuracy.

Also, a slightly differant circuit for the Video Sync buffer circuit..


I've not tested the circuit but should be ok... (hopefully).

Expanding your circuit, I have this so far (I've added the luma line, but I suppose that's not necessary for a CM14). I was wondering why you have a different circuit for the sync buffer, and if you have any ideas about the audio? I'm looking at audio splitter circuits and they're all surprisingly complicated... Most seem to use a particular opamp and require 12V input... Ideally, I'd like to keep to just 5V input and using basic components... Any suggestions?

[RetroCPC]

Hi Chris,


I'd also add some High Frequency decoupling on the 5V Power Rail to Ground.... say 4x 100nF ceramic (spread between the buffer stages) + say 100uF Electrylitic "Bulk" Capacitor.


I tired to keep the Sync Signal output impedance the same as the CPC output (after buffering) - it might not be needed, but as I have no moniter to test, I wanted to play safe.


Audio is more difficult as you require signals swinging around 0V - (symetrical Voltage swing around Zero volts). Obvusally you can AC couple - but with 5V rails you really require Rail to Rail circuitry... Sure it can be done, but my heart aches to design something of poor quality....

Give me 30 minutes and I'll design something semi-decent for you....

[RetroCPC]

Hi Chris,

Heres a descrete Output Audio Buffer, you will need one "Pair" for each audio channel (L/R)

It might be too much for your needs, but will work well, and cheap components... however you will have fun designing the PCB...

Clean 5V power would be very helpful, as Power supply rejection is rather limited (40dB at 50Hz, 66dB at 1KHz & 92dB @ 20KHz)... if its a problem, then C6 can be increased to improve the noise rejection...

The Electrolytic Capacitors need to be 6.3V or greater...

[Cwiiis]

Hi Chris,

Heres a descrete Output Audio Buffer, you will need one "Pair" for each audio channel (L/R)

It might be too much for your needs, but will work well, and cheap components... however you will have fun designing the PCB...

Clean 5V power would be very helpful, as Power supply rejection is rather limited (40dB at 50Hz, 66dB at 1KHz & 92dB @ 20KHz)... if its a problem, then C6 can be increased to improve the noise rejection...

The Electrolytic Capacitors need to be 6.3V or greater...

Seems the audio part of this equations is oddly much more complicated than the video part I've attached the circuit I have so far, but I'm not sure I really understood your suggestion with the decoupling capacitors - I have the power input smoothing capacitor there (100u seems high?), but I don't know that I've put the 4 decoupling capacitors you suggested in the right place... That said, that's the only place where there are 4 components that match up, so maybe?

Re the audio, I found this nice, simple audio splitter circuit: https://www.learningelectronics.net/circuits/laptop-audio-out-splitter-circuit.html - it uses a couple of TDA2822M ICs, but I think the trade-off is very much worth it... I suppose using this circuit, we could replace P1/P2 with resistors (though not sure what value... I suppose I could find this out by using a potentiometer and comparing with straight-through, or perhaps just providing voltage to it and seeing what comes out?)

[Bryce]

Video signals are ground referenced and vary between 0V up to 2 or 3V. Audio signals have the ground reference in the middle, ie: they vary from a +V to -V which makes the electronics slightly more complicated. RetroCPC's circuit looks complicated because it's a purely discrete circuit, ie: made up of transistors and passive parts. If this is too complicated, you could always use the circuit that Amstrad used for the internal audio, however, RetroCPC'S circuit will give you a much (much, much....) better audio quality than the Amstrad circuit.

Bryce. 

[RetroCPC]

I've attached the circuit I have so far, but I'm not sure I really understood your suggestion with the decoupling capacitors - I have the power input smoothing capacitor there (100u seems high?), but I don't know that I've put the 4 decoupling capacitors you suggested in the right place... That said, that's the only place where there are 4 components that match up, so maybe?

Chris,

As you suspected, the 100nF capacitors I mentioed are in the incorrect location on your schematic, they need to be connected between the +5V rail and Ground - the same as you have the 100uF Cap connected.

I was imagining them evenly spread across the PCB between sets of the video buffer transistors. There function is to offer high frequency PSU decoupling... I'm sure the circuit would work without them, but better with them.

The TDA2822M circuit you linked to is a mini power amplifer (to drive small speakers - think those cheap USB powered speakers) and headphones rather then a audio splitter for sure it could still work as a splitter... (but you dont really want any Gain) - and typically power amplifers have 25dB to 30dB Gain (the TDA2822M has 39dB fixed Gain!!)...

You mnentioned you wanted to use "using basic components" so I tock this as being "basic" discrete components...

[Cwiiis]

Chris,

As you suspected, the 100nF capacitors I mentioed are in the incorrect location on your schematic, they need to be connected between the +5V rail and Ground - the same as you have the 100uF Cap connected.

I was imagining them evenly spread across the PCB between sets of the video buffer transistors. There function is to offer high frequency PSU decoupling... I'm sure the circuit would work without them, but better with them.

The TDA2822M circuit you linked to is a mini power amplifer (to drive small speakers - think those cheap USB powered speakers) and headphones rather then a audio splitter for sure it could still work as a splitter... (but you dont really want any Gain) - and typically power amplifers have 25dB to 30dB Gain (the TDA2822M has 39dB fixed Gain!!)...

You mnentioned you wanted to use "using basic components" so I tock this as being "basic" discrete components...

Ah, I think I might understand what you mean with the capacitors - I've moved them to the video lines coming off Q2, Q4, Q6 and Q8.

Your assumption of what I meant was correct, I just didn't imagine it to end up being quite so densely populated I suppose this is what it takes though. Please do excuse my naivete, hardware/electronics really isn't my field... I can't seem to find a simple IC that takes a line-level audio input and gives 2 line-level audio outputs (assuming the Amstrad is even line level? Or somewhere close?), which would be the ideal (well, 2 in 4 out would be the ideal I suppose), so I'll go ahead and integrate your design. I'll test the video circuit first as I have most of those parts already, the audio may have to wait a bit.

[RetroCPC]

Hi Chris,


Can you forward your updated schematic just to confirm you have the capacitors connected correctly, they should be connected between +5V and Gnd - not to any of the Video / Sync signals.

[Cwiiis]

Hi Chris,


Can you forward your updated schematic just to confirm you have the capacitors connected correctly, they should be connected between +5V and Gnd - not to any of the Video / Sync signals.

I had it one way, but I think maybe I realise what you were saying now and have it like this. Close?

[RetroCPC]

Hi Chris,


Yes, looks good now - you can drop C6 as you have plenty of decoupling caps now

I'm kind of hoping you will use Ground plane (Copper Fill for the GND signal)


If you use skype then I can talk you though the PCB layout - I'd do it for you but I'm fighting my own PCB Layout ATM.... or I should say more correctly its fighting me a WINNING!!!

[Cwiiis]

Hi Chris,


Yes, looks good now - you can drop C6 as you have plenty of decoupling caps now

I'm kind of hoping you will use Ground plane (Copper Fill for the GND signal)


If you use skype then I can talk you though the PCB layout - I'd do it for you but I'm fighting my own PCB Layout ATM.... or I should say more correctly its fighting me a WINNING!!!

I very much appreciate the offer! I'm going to test on breadboard first, see if the concept will work - if it does, maybe I'll make a video-only version (well, one output will have audio!) on veroboard or something. I've done PCB layout before, so I can give it a go, but I'll definitely be consulting with you if you have the time

I like your audio circuit, but I'm considering 4xLME49721 as buffers, mainly to make layout a bit easier. Either way, first things first, I'll think about audio when video is working. Missing parts should be arriving tomorrow, so I'll update once I get the time to try it out. Thanks for all your help so far!

[Cwiiis]

I've tested the circuit out on breadboard now (took a while to get the parts I didn't have) and it appears to work! Just thought I'd update here in case this it's useful for anyone else - I'm going to go ahead and design a PCB now. I haven't gotten the parts to test audio splitting with the opamp yet, so I'm going to rough out a circuit and if it's wrong I can just break out audio separately (or patch it through temporarily to just one of the outputs). I'll update again once I have a design.

[Cwiiis]

Hi Chris, Yes, looks good now - you can drop C6 as you have plenty of decoupling caps now I'm kind of hoping you will use Ground plane (Copper Fill for the GND signal) If you use skype then I can talk you though the PCB layout - I'd do it for you but I'm fighting my own PCB Layout ATM.... or I should say more correctly its fighting me a WINNING!!!

Attached is the updated circuit with the audio splitter and preliminary PCB layout.I've used a ground plane, but only on the top layer as everything is through-hole except for the two op-amps which are soldered on the top of the board. This could be a much smaller board but I figure it's nice to have a bit of space to work with and possibly do any necessary bodging Any thoughts/feedback? I'm very uncertain of the values around the audio circuitry, but I guess that's what I get by diverging from your design!

[Bryce]

Attached is the updated circuit with the audio splitter and preliminary PCB layout.I've used a ground plane, but only on the top layer as everything is through-hole except for the two op-amps which are soldered on the top of the board. This could be a much smaller board but I figure it's nice to have a bit of space to work with and possibly do any necessary bodging Any thoughts/feedback? I'm very uncertain of the values around the audio circuitry, but I guess that's what I get by diverging from your design!

Looks good, but PCB space costs money, you could easily half the size and still have room for bodges etc.

Bryce.

[Cwiiis]

Looks good, but PCB space costs money, you could easily half the size and still have room for bodges etc. Bryce.

You're right - I've reduced the empty space somewhat to bring it down to an outline of 126x90mm. It could still be much smaller, but I think this is a reasonable compromise - I'm optimising more for putting it together than I am for size/cost, I can't imagine there are too many people out there looking to split their analog Plus output

If I was planning on making a few, or I wanted to make a miniature/cheaper version, I'd replace all of these components with their 0603 surface mount equivalents and have the board be barely bigger than the connectors - but I'll leave that to someone else for now!

[Bryce]

You could also put dual footprints, for the equally low number of people who would want to split VGA signals? 

Bryce.

[RetroCPC]

Any thoughts/feedback? I'm very uncertain of the values around the audio circuitry, but I guess that's what I get by diverging from your design!

Hi Chris,

You need to AC couple and bias the input of the Opamp buffers, I've attached a quick schematic (you need to power the opamp from the 5V rails - the Power supply and Ground connections to the opamp are not indicated on my schematic).

I've also added R1, R9, R11, R7, R8, R10 (these need to be located as close as possible to the opamp) - there function is to insure the opamp remains stable (does not oscillate) and also some input RF filtering (C4).

I'd Ground fill both layers - and liberally add Vias spread around the PCB to connect the Top and Bottom Ground planes...

[Cwiiis]

Hi Chris,

You need to AC couple and bias the input of the Opamp buffers, I've attached a quick schematic (you need to power the opamp from the 5V rails - the Power supply and Ground connections to the opamp are not indicated on my schematic).

I've also added R1, R9, R11, R7, R8, R10 (these need to be located as close as possible to the opamp) - there function is to insure the opamp remains stable (does not oscillate) and also some input RF filtering (C4).

I'd Ground fill both layers - and liberally add Vias spread around the PCB to connect the Top and Bottom Ground planes...

Thanks again, this kind of circuit is definitely beyond me right now... I guess the extra resistors essentially drive the signal? Is C3 definitely the right way round? It looks like it would just block input, but maybe there's something I don't understand there... Obviously, any explanation is a huge courtesy, but I'd very much appreciate it

[RetroCPC]

Chris,

Your power connector is rotated the wrong way around - the PCB mounting pads are at the rear of the connector (not the front) - as you have it currently, the power input (Hole) is facing towards the center of the PCB - not the edge of the board....

The Opamp needs to be biased to half of your PSU supply (so 2.5V) - ideadly the audio input should swing about 0V (but the Amstrad is not AC coupled) - anyway, the Capacitor C3 has 2.5Vdc on the Opamp side (biased by R2/R3/R5 - with C1 offering some PSU noise filtering) and hopefully something less (idealy swinging about 0V) on the Audio input side...

The circuit uses the same LME49721 you use - but with a few extra components I can confirm your PCB layout if you give it a try.


Also, I'd move your Audio input signals away from the Video pin connector as the you will pickup a "Buzz/ hum" from the Sync signals / Video data...

[Cwiiis]

Chris,

Your power connector is rotated the wrong way around - the PCB mounting pads are at the rear of the connector (not the front) - as you have it currently, the power input (Hole) is facing towards the center of the PCB - not the edge of the board....

The Opamp needs to be biased to half of your PSU supply (so 2.5V) - ideadly the audio input should swing about 0V (but the Amstrad is not AC coupled) - anyway, the Capacitor C3 has 2.5Vdc on the Opamp side (biased by R2/R3/R5 - with C1 offering some PSU noise filtering) and hopefully something less (idealy swinging about 0V) on the Audio input side...

The circuit uses the same LME49721 you use - but with a few extra components I can confirm your PCB layout if you give it a try.


Also, I'd move your Audio input signals away from the Video pin connector as the you will pickup a "Buzz/ hum" from the Sync signals / Video data...

God, I can't believe I didn't notice the power connector, thanks for that! 




Thanks for the explanation and more tips, I'm revising the schematic and PCB now, I'll update again with the next revision - and hopefully the correct one Re the video pins causing audio interference, I suppose the individual cables coming off the DIN connector ought to be shielded in that case?

[Cwiiis]

Hi Chris,

You need to AC couple and bias the input of the Opamp buffers, I've attached a quick schematic (you need to power the opamp from the 5V rails - the Power supply and Ground connections to the opamp are not indicated on my schematic).

I've also added R1, R9, R11, R7, R8, R10 (these need to be located as close as possible to the opamp) - there function is to insure the opamp remains stable (does not oscillate) and also some input RF filtering (C4).

I'd Ground fill both layers - and liberally add Vias spread around the PCB to connect the Top and Bottom Ground planes...

Something I'm uncertain about in this circuit (amongst other things ) is how to integrate C5 and C6 with the larger circuit - I suppose this is a combination of decoupling/smoothing(?), but I already have C1 (though it's a much smaller value) for smoothing and C10 and C11 for decoupling, and I'm not sure how to go about duplicating this for two sources, or its intention. I've attached what I have so far, which I think is the whole circuit minus C5 and C6 from your diagram and with power input/grounds for the opamps with decoupling on C10 and C11.