ZX81 rebuild

Discussions about Sinclair ZX80 and ZX81 Hardware
Lardo Boffin
Posts: 672
Joined: Sat Nov 26, 2016 1:42 am

Re: ZX81 rebuild

Post by Lardo Boffin » Fri Dec 01, 2017 10:55 am

1024MAK wrote:
Fri Dec 01, 2017 12:11 am
What type of transistor is fitted in TR2 position?

Mark
It is a new old stock Motorola MPS2369. Transistor tester has been ordered and is on its way...
ZX81 iss 1 (bugged ROM, kludge fix, normal, kit)
Iss 3 with ZXVid, another in nearly perfect box
TS 1000 iss 3, ZXPand AY, ZX8-CCB & ZX-KDLX
TS 1500
TS 2000
Spectrum 16k (iss 1 s/n 1499)
Spectrum 48k
Toastrack + DIVMMC future

Lardo Boffin
Posts: 672
Joined: Sat Nov 26, 2016 1:42 am

Re: ZX81 rebuild

Post by Lardo Boffin » Fri Dec 01, 2017 11:06 am

1024MAK wrote:
Thu Nov 30, 2017 11:36 pm

Check / test R5, R6, C7 and TR2.
R3 should not be fitted.

Mark
Looking down on the zeddy from above

R3 is not fitted
R5 voltage left = 4.98, right = 0.11
R6 voltage left = 1.65, right = 0.75

Not sure how you test a capacitor (its brand new and I double checked it is the correct 47pF rating) but
C7 voltage left = 1.62, right = 0.72

I will check TR2 when the tester arrives but given that the other items involved are brand new I suspect this is the most likely cause of issues. I may well steal one from a working zeddy and see if it helps...
ZX81 iss 1 (bugged ROM, kludge fix, normal, kit)
Iss 3 with ZXVid, another in nearly perfect box
TS 1000 iss 3, ZXPand AY, ZX8-CCB & ZX-KDLX
TS 1500
TS 2000
Spectrum 16k (iss 1 s/n 1499)
Spectrum 48k
Toastrack + DIVMMC future

User avatar
1024MAK
Posts: 1821
Joined: Mon Sep 26, 2011 9:56 am
Location: Looking forward to summer in Somerset, UK...

Re: ZX81 rebuild

Post by 1024MAK » Fri Dec 01, 2017 11:40 am

Sorry, I should have been more specific, but in my defence, at the time of writing my post, a two carriage train full of people got squashed into the rear carriage because the driver and the guard locked the front carriage out of use due to one set of defective power operated doors.

With the power disconnected, use the resistance range on your meter to test the resistance of R5 and R6.
Check that capacitor C7 is indeed 47pF. Check that the transistor is correctly fitted, as in check the pin-out (lead-out) of the transistor on the data sheet is correct to the actual PCB connections.

No need for a transistor tester, from your results, TR2 is "on", as it is conducting current from it's collector to it's emitter. The problem is that it is not switching off when it should.

In this circuit, the transistor type is important, and in some circumstances, the value of R6 may need to be changed to suit the type of transistor used for TR2. I have a Zeddy with a MPS2369 transistor somewhere. So hang fire while I find it...

Mark

Lardo Boffin
Posts: 672
Joined: Sat Nov 26, 2016 1:42 am

Re: ZX81 rebuild

Post by Lardo Boffin » Fri Dec 01, 2017 12:02 pm

No need for an ‘in your defence!’ :oops: I need one though...

R5 323 ohm
R6 2.15 k ohm

So both fairly close to the required settings.
ZX81 iss 1 (bugged ROM, kludge fix, normal, kit)
Iss 3 with ZXVid, another in nearly perfect box
TS 1000 iss 3, ZXPand AY, ZX8-CCB & ZX-KDLX
TS 1500
TS 2000
Spectrum 16k (iss 1 s/n 1499)
Spectrum 48k
Toastrack + DIVMMC future

User avatar
1024MAK
Posts: 1821
Joined: Mon Sep 26, 2011 9:56 am
Location: Looking forward to summer in Somerset, UK...

Re: ZX81 rebuild

Post by 1024MAK » Sun Dec 03, 2017 8:41 pm

I've found a Zeddy that uses a MPS2369 transistor in the TR2 position. The values of the resistors are the same as those you are using. It's an issue 3 board.

As I said in my earlier post, TR2 is not switching off. So assuming it is correctly connected, you could try desoldering one end of R6, then soldering in a temporary wire. You can then connect test resistors (in series) between the now unconnected lead of R6 and the temporary wire. The total resistance will be the sum of R6 plus whatever value of your test resistor. We need to increase the total resistance so that TR2 is driven with a lower current, so that it can switch off quicker. When it's base current is low enough, it will switch off when the clock output from the ULA is low. Then the clock input pin to the Z80 (pin 6) will go high.

Good luck ;)

Mark

User avatar
1024MAK
Posts: 1821
Joined: Mon Sep 26, 2011 9:56 am
Location: Looking forward to summer in Somerset, UK...

Re: ZX81 rebuild

Post by 1024MAK » Sun Dec 03, 2017 9:06 pm

Photos of TR2 and R6 on the issue 3 board:-
IMG_5806.JPG
TR2 on an issue 3 board
IMG_5806.JPG (805.04 KiB) Viewed 217 times
IMG_5805.JPG
R6 on an issue 3 board
IMG_5805.JPG (696.9 KiB) Viewed 217 times
Mark

Lardo Boffin
Posts: 672
Joined: Sat Nov 26, 2016 1:42 am

Re: ZX81 rebuild

Post by Lardo Boffin » Sun Dec 03, 2017 10:03 pm

Thanks Mark. I will give that a go. Below is how my TR2 is connected:-

03E20040-0F7A-48A7-AA6F-A84A0D4C884D.jpeg
03E20040-0F7A-48A7-AA6F-A84A0D4C884D.jpeg (658.64 KiB) Viewed 205 times
ZX81 iss 1 (bugged ROM, kludge fix, normal, kit)
Iss 3 with ZXVid, another in nearly perfect box
TS 1000 iss 3, ZXPand AY, ZX8-CCB & ZX-KDLX
TS 1500
TS 2000
Spectrum 16k (iss 1 s/n 1499)
Spectrum 48k
Toastrack + DIVMMC future

Lardo Boffin
Posts: 672
Joined: Sat Nov 26, 2016 1:42 am

Re: ZX81 rebuild

Post by Lardo Boffin » Mon Dec 04, 2017 1:03 pm

Prior to experimenting with resistors I quickly swapped out the MS2369 and put back in the original transistor and all of a sudden...

DF1AA09F-0C9D-496F-8A9D-77F6C4CB27E6.jpeg
DF1AA09F-0C9D-496F-8A9D-77F6C4CB27E6.jpeg (492.12 KiB) Viewed 189 times

So it is definitely the 2369 causing problems and amazingly everything else is Ok!

As soon as I get some spare time I will put the 2369 back in, mess with resistors and finish the build! Whu hu! 8-)

(For anybody wondering why in earth I didn’t just do that in the first instance the whole point of this was to rebuild a zeddy with as few original components as possible so that it feels like my computer not a second hand one :D )
ZX81 iss 1 (bugged ROM, kludge fix, normal, kit)
Iss 3 with ZXVid, another in nearly perfect box
TS 1000 iss 3, ZXPand AY, ZX8-CCB & ZX-KDLX
TS 1500
TS 2000
Spectrum 16k (iss 1 s/n 1499)
Spectrum 48k
Toastrack + DIVMMC future

Lardo Boffin
Posts: 672
Joined: Sat Nov 26, 2016 1:42 am

Re: ZX81 rebuild

Post by Lardo Boffin » Tue Dec 05, 2017 1:33 pm

I started off adding 100ohm resistors but got nowhere so swapped for 370ohm and finally got to about 4k4 at which point it pulses but the Lo light stays on as well. If I add another 370 it still pulses but the Hi light comes on instead. Is there a ‘sweet spot’ inbetween or is 4k4 going to be ok? Would certainly be the easiest!

Thanks

Lardo
ZX81 iss 1 (bugged ROM, kludge fix, normal, kit)
Iss 3 with ZXVid, another in nearly perfect box
TS 1000 iss 3, ZXPand AY, ZX8-CCB & ZX-KDLX
TS 1500
TS 2000
Spectrum 16k (iss 1 s/n 1499)
Spectrum 48k
Toastrack + DIVMMC future

User avatar
1024MAK
Posts: 1821
Joined: Mon Sep 26, 2011 9:56 am
Location: Looking forward to summer in Somerset, UK...

Re: ZX81 rebuild

Post by 1024MAK » Tue Dec 05, 2017 5:31 pm

The reason that Sinclair chose a ZTX313 is that it has a very fast switch off time. The MPS2369 transistors are aimed at video amplifiers and similar HF applications. So although they are fast, they may not switch off as quickly as a ZTX313.
Lardo Boffin wrote:
Tue Dec 05, 2017 1:33 pm
I started off adding 100ohm resistors but got nowhere so swapped for 370ohm and finally got to about 4k4 at which point it pulses but the Lo light stays on as well. If I add another 370 it still pulses but the Hi light comes on instead. Is there a ‘sweet spot’ inbetween or is 4k4 going to be ok? Would certainly be the easiest!
The trick is to find a base resistor that allows just enough current to the transistors base (input) so that it switches on enough that the voltage at the collector (output) falls to zero (logic zero), but without turning on the transistor too hard. If there is more current flowing into the base than needed, the transistor takes longer to switch off.

And just to complicate matters, transistor gain will vary between individual transistors and also slightly with temperature variations.

Without an oscilloscope, it's a matter of finding a value where the Z80 CPU works. In an ideal world, the logic probe's hi and lo LEDs would be both on and at the same brightness, but that also partly depends on the model of logic probe.

So if 4.4k works, leave it in circuit and see how reliable the computer is.

If I get time (next week), I will hook a 'scope up to the ZX81 that I have that uses MPS2369. I'll post pictures of the waveform as well.

Mark

Post Reply

Who is online

Users browsing this forum: No registered users and 5 guests