Gilding the DKTronics Speech Synthesizer

[CanonMan]

Thank you for your concise reply. Good question about the Motherboard - I have no idea as I've just purchased it from EBay - sold as Non Working. The seller had replaced the drive belt - but it was still unable to read the disks he had onhand.

Replacing the drive belt isn't always enough on its own.


At the very least, you should make sure the pulleys have been cleaned (some people skip this!) and check all the PCB connector pins for dry joints - it's quite common.

[RetroCPC]

CanonMan,

I'll take a look at the pulleys and clean them. There where a few doggy looking solder joints especially on the connectors - some of the main QFP60 controller IC joints look alittle suspect so I'll also go over these - many of the IC's legs don't appear to be wetted during the Wave soldering process, although they dont appear to move when poked...

While I was rummaging around I found one of those tiny I2C OLED modules that can be retro fitted to the Gotek USB drive emulators so I might give that a try.. its a question of finding the time...

[RetroCPC]

Just a quick update WRT to my CPC advantages and getting to work on a few CPC hardware designs:-

I decided to bypass my CPC FDD issues and I just ordered a "AMSTRAD CPC 6128 - GOTEK - 3D PRINTED MOUNT - OLED DISPLAY - FLASH FLOPPY" from Ebay, This also solves how to get software into my CPC in the first place (I understand I can just download .DSK files from the internet onto a USB flash drive and then plug into the Gotek "Drive").

Second issue is that I dont have a suitable monitor - so I ordered BenQ BL702A 17.0 inch LED Monitor which is meant to work the CPC's "PAL" 15KHz frame rate - the idea is to modify my CPC with a "VGA port" - tapping the HSync / V Sync signals from the 6845 CRT controller and using the standard RGB signals... I'm not sure if this will work... but if it does I hope this will be the simplest solution - with no extra signal loss / degradation from an external Scan rate converter.

If this "VGA" modification  does not work out, then as a backup plan I've also ordered a "Gonbes GBS-8200 CGA (15kHz)/EGA (25kHz)/YUV/RGBS to VGA HD Video Converter 8200" again from Ebay... Apparently this PCB can be modified with an external MCU to optimize the boards settings for the CPC...

Shame there does not appear to be a simple display solution for the CPC - as someone whose just getting back into the CPC scene since 1980's this is being my biggest headache... (I'd rather not go back to a chunky / hot CRT monitor)...

[VintageAdvantage]

(I'd rather not go back to a chunky / hot CRT monitor)...

That's what I thought too when I started again. Had (and still have, for the most of it) hooked everything up to a TFT with a converter / upscaler.

However, CRTs add A LOT to the retro experience. Much more fun with a CRT... so maybe you'll change your mind one day again 

[RetroCPC]

I'm tempted into looking into designing a true digital monitor interface for the CPC - with no conversation from Analogue to digital... this will give a pixel accurate display, my eyes are very sensitive to fizzy pixels...

Windows 10 display scaling is really poor - WindowsXP with its pixel perfect display is so much clearer to my eyes - especially when working on a PCB layout etc...


WRT to a CRT monitor, my little man cave is 5m x 5m (external dimensions) I simply have no space for a CRT monitor... I have 4 dogs who are attached to my every move... they insist on each being in there own baskets on my already limited bench space!!! so I work sandwiched between dog baskets... a CRT monitor is not happening!

[Audronic]

Yes be tempted


Ray

[RetroCPC]

Yes be tempted

Ray

Thinking of a small PCB that intercepts the signals to the Amstrad ASIC and maybe the CRT controller (for access to timing) - maybe this timing info (Pixel clock) could be recreated in the FPGA / MCU... as I lack VHDL skills, I'd rather use an MCU then FPGA (although this is a really a task for a small FPGA, something like an ICE40 / 50) - maybe I could use a smaller XMOS... it be a way to hone some XMOS programming  skills....


I'd need to first understand how the CPC maps the colour info into memory...  luckily we only have 1.5 bits, 0, 50% and 100% for each RGB colour...

[Bryce]

I use a Commodore 1084 CRT with SCART input. It's obviously deeper than an LCD monitor, but doesn't really take up more space otherwise.

Bryce.

[RetroCPC]

Bryce,

I've had time to wire up a VGA connector + rats nested a buffer to provide the H/V sync from the 6845 CRT controller and now have a stable image on the BenQ LCD monitor but the colors are wrong.

The first powerup text is Yellow but with a very dark blue background (almost black really) - not the typical lighter blue background as seen on the CRT - also on Oh Mummy the expected Yellow background is washed out very dark yellow - with black ghosting edges, could be some kind of video overload... Odd as the text on powerup looks OK (Yellow)

I suspect I have either issues with termination or incorrect video signal levels for a VGA monitor...

Do you have any hints suggestions?

I'm not sure of the correct video RGB levels / termination required for the CPC - maybe I should more correctly say for VGA...

[Bryce]

What circuit are you using for the buffer?

Bryce.

[RetroCPC]

What circuit are you using for the buffer?

Bryce.

I'm only logic buffering the H/V Sync from the 6845 (via a couple of 74HC04 gates connected in series for each Sync signal)... the RGB signals are taken directly from the CPC's Din socket - I've since read that the signal amplitude for VGA is 700mV terminated into 75ohms (150ohms Zout from the source, 150 ohms @ the VGA monitor).

I need to take a closer look at the CPC video circuit...

Found this in the hardware manual:

The input will be the same as the earlier GTM65 versions, i.e. impedance 470 ohms to 0V, analogue voltage input which is linear between 0.8V (Black) and 1.75V (Peak white).
The CM14 colour monitor needs to handle a sixteen level input on each of RGB. The new monitor must present an input impedance of 100 ohms to 0V.

This does suggest I need to pad down the signal to 700mV for VGA, I need to scope the signals to better understand whats going on...

[RetroCPC]

More info on VGA video levels and termination:

Distributing the near end termination resistance is an attempt to reduce reflections at the VGA connector.  However, it's sloppy.  Looking from the VGA output, the nearby resistor looks like nothing more than a resistor, and can be ignored.  Then you see 50 ohm line, then 150 ohms in parallel with the 75 ohm coax and far end termination at the monitor.  150 ohms at the connector in parallel with 75 ohms = 50 ohms.  The 50 ohm line is matched and does not "see" the transition into the 75 ohm line. 

The nominal video level for a VGA monitor is .7 VPP, with black at ground.

All is well if the monitor itself is well terminated.  This is not always the case.  Reflections coming back on the 75 ohm line encounter first, 150 ohms paralleled by the 50 ohm line, or 37.5 ohms, yielding a nasty negative re-reflection.  However, at the far end of the 50 ohm line, the signal will meet the 150 termination at the chip, which makes another nasty but positive reflection.  Keeping the distance between these mismatches short helps to overlap the reflections to achieve some cancellation.  The time skew of several ns between them leads to spikes of noncancellation traveling back to the monitor that are hopefully beyond the monitor.

[RetroCPC]

Scoped the Video signals to the monitor, and levels appear correct, but there is so much noise (especially Ground bounce) &  reflections as a result of poor termination, signal integrity is a real mess.

I think one of the problems is the higher video B/W of high resolution monitors is just magnifying the CPC's poor by today standards video output...

The CPC6128 lacks a Ground plane and the ULA the produces the video signal is a distance away from the video output connector resulting in long incorrectly impedance matched RGA PCB tracks + huge ground noise between the ULA ground pins and video output connector...

Pondering moving the ULA, clock circuit and video interface to a small 4 layer daughter board with improved layout to see if this helps to improve video quality...

Attached is a relativity correct photo of the LCD screen  - really odd dark horizontal bars and noise is clearly visible... the vertical bar to the left of the picture is just a result of underscan, the BenQ BL702A monitor cannot overscan - despite supporting 15KHz, its not a great solution...

[RetroCPC]

Spent more time then I should have understanding the CPC 40010 Gate Array video out.

In the end its very simple - and has 3 states:-

100%: Active pulled high via a very non linear "resistor" I suspect its more likely simply the RdsOn of the Pch device. Measures 75 Ohms when Lightly loaded, nearer 100 Ohms when loaded.


50%: this is Tristate (no active drive) - so the voltage is determined by external resistor values, this allows one to trim the colours.


0%: Active low, pulled down via ~ 22 ohms.


There is significant spectral noise on the GateArray outputs - with significant ground bounce, Jitter and other modulation, adding to the poor signal integrity is the PCB's layout  lack of any ground plane and controlled impedance's which significantly worsens the video signal.

Without video buffering (which will not help with the spectral quality / noise) I can see no way of achieving a fixed output impedance (and thus correctly terminated signals) to drive a VGA monitor (or really any monitor if we where being pedantic).

Now that I understand the GateArrays output stage, I can design an adapter PCB to provide a correct impedance / signal level, low phase noise video output.

[VintageAdvantage]

Is the DKtronics done yet?

[RetroCPC]

I got sidetracked working on improving the CPC6128 Video quality (I now have a decent picture quality) - then my day job got in the way...

Still plan a Gilded Speech Synth / AmDrum PCB just for the fun of it...

I'll get back to it when I complete my current PCB design...