ZX Spectrum +2 booting problem

For experts to discuss very technical stuff and newbies to ask why the Spectrum they bought off ebay doesn't work.
User avatar
1024MAK
Bugaboo
Posts: 3104
Joined: Wed Nov 15, 2017 2:52 pm
Location: Sunny Somerset in the U.K. in Europe

Re: ZX Spectrum +2 booting problem

Post by 1024MAK »

You could do that. Keep in mind however that unlike the 4116 DRAM chips fitted in the 16k and 48k machines which are known for sometimes failing. The TEA2000 chips are normally fairly reliable. That does not mean that it has not failed. The +12V and -12V supply also goes to IC3 (1488) which is used for level shifting to RS232 voltages (for the serial and aux/keypad ports). +12V also goes to various resistors and to TR4.

If you don’t have a current limited supply for the +12V, you could use any regulated PSU, but wire a 100 ohm resistor between the +12V output of the PSU and the +12V rail on the ZX Spectrum +2. Then you can measure the voltage across this 100 ohm resistor. The resistor will also limit the current if there is a short circuit (it will get hot in this case).

Voltage (U) in volts (V) divided by resistance (R) in ohms gives the current (I) in Amps (A)

So if you get 2.5V...

Code: Select all

U       2.5V
— = I = ———— = 0.025A = 25mA
R       100
This method gets increasingly impractical as the current increases in a design. As more power is lost in the resistor, and you also have the problem of the voltage ‘lost’ across the resistor means that the circuit that is being supplied no longer sees the correct supply voltage.

There are other ways of working out where short circuits are, but in practice, on printed circuit boards, most are impractical. In order to measure current using a multimeter, the meter needs to be connected in series, so the current flows through it. This means breaking one of the conductors of the circuit you want to test.

The only other ways (without disturbing the PCB air the components) are to supply the circuit with power and either use a very sensitive voltmeter to compare the chip pin voltages with a known good unit. Or use an expensive IR camera to detect components that are warmer than they should be.

Mark
:!: Standby alert :!:
“There are four lights!”
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :dance
Looking forward to summer later in the year.
CapSmasher
Drutt
Posts: 16
Joined: Mon Oct 22, 2018 9:53 am

Re: ZX Spectrum +2 booting problem

Post by CapSmasher »

Thank you very much for your answer...

Sorry for asking all these questions that the answer can be pretty obvious for an experimented person. Is clearly making some light in the understanding of how these circuits work and hope that will also help others.

So basically I could first try checking if TR4 (2N3904) is not the short's cause, then the TEA2000? If I exclude these there will be only a matter of a failed resistor/capacitor in the way - or that IC3 (MC1488P)?

Thank you
CapSmasher
Drutt
Posts: 16
Joined: Mon Oct 22, 2018 9:53 am

Re: ZX Spectrum +2 booting problem

Post by CapSmasher »

Following the post above,
TR4 seems to work... Maybe I should try pumping 12v + 120 Ohm (0.1A) resistor to the D20 diode and check some voltages? Do you have some advice what voltages should I check first to spot this cursed short?
User avatar
1024MAK
Bugaboo
Posts: 3104
Joined: Wed Nov 15, 2017 2:52 pm
Location: Sunny Somerset in the U.K. in Europe

Re: ZX Spectrum +2 booting problem

Post by 1024MAK »

CapSmasher wrote: Tue Nov 06, 2018 9:00 am TR4 seems to work... Maybe I should try pumping 12v + 120 Ohm (0.1A) resistor to the D20 diode and check some voltages? Do you have some advice what voltages should I check first to spot this cursed short?
It’s really a case of trying it, and then seeing what the results are.

This test will may not tell you what is causing the short, but each step narrows down the possibilities.

Mark
:!: Standby alert :!:
“There are four lights!”
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :dance
Looking forward to summer later in the year.
CapSmasher
Drutt
Posts: 16
Joined: Mon Oct 22, 2018 9:53 am

Re: ZX Spectrum +2 booting problem

Post by CapSmasher »

Ok, so I tried with some primitive PSUs only for the 5v line. I connected the positive terminal from the power source to the positive pin of the voltage regulator output, and the negative to the negative pin nearby.

I observed that I have some massive voltage drop - somewhere like 4.1v are reaching the RAM chips - it would be wise to find a power source of let's say 6v and test to see if ~5v would reach the chips? Or there is something on the board that cause this voltage drop. The power source alone had at least 5.1v.. Another observation was (with another -new- multimeter) that when I first touched the RAM power pin it showed 5. something volts, but then instantly dropped at 4.1v

I also put a resistor of 10 Ohm and 100 Ohm, and indeed they are getting hot pretty fast (the 10 Ohm one got it to the point that it smell, but I saved it in the last moment). In the end I saw that the power source was not going more then 1.2A and I removed the resistor . Interesting is that the TEA2000 have the same voltage as it was powered before (of 1.2v something) that means the +12v line was dead and this 1.2 voltage is in fact an 'echo' from the 5v line !? I think that using the resistors also drastically increase the voltage drop to something like 2.5v

It is strange that this voltage drop is higher than expected, with the voltage reglator when I had 5v output, on RAM chips I had somewhere like 4.7v...
CapSmasher
Drutt
Posts: 16
Joined: Mon Oct 22, 2018 9:53 am

Re: ZX Spectrum +2 booting problem

Post by CapSmasher »

This method gets increasingly impractical as the current increases in a design. As more power is lost in the resistor, and you also have the problem of the voltage ‘lost’ across the resistor means that the circuit that is being supplied no longer sees the correct supply voltage.
I just read this again and makes sense, but I had high voltage drop also without using the resistor...

In the meantime I found that the Composite Sync is using 1.2v and that is just the current value on TEA2000's pin11. I guess it could be a sign that TEA2000 may be functional? If is true, is curious that it receives 1.2v without any loss , compared to the ram chips. The only chip that was getting warm was the CPU.

Image

Ok maybe the plan with using an external source was too ambitious with my current tech and experience...

What if I'll do like this:

1. connect the Spectrum as usual from a 48k power supply - will be safe to leave TR9 out just for testing 5v functionality ?
2. improvise a wiring from the RGB port to some SCART jack (don't have a cable for this) - do you think bare wires will work for this or I have to put some resistors on R,G,B?
Image

3. observe if something is going on screen

I'm still a bit surprised that I had +12v dead at some point but still had some fuzzy image with analog output (when voltage regulator gone berserk). I remember that I still had something like 1. something volts on TEA2000 so in theory analog video should be dead?

LATER EDIT:
I just read on some comments on other sources that the Serial Transceiver Chip (I think is about that MC1488P ?) is common to kill the +12v line. It is safe to remove this chip (or just some pins) just to test if is indeed the short's cause ?

EVEN LATER EDIT:
I got my hands on some TIP31C (to replace ZTX653 / TR9), do you think this could handle better the short and survive the tests?
CapSmasher
Drutt
Posts: 16
Joined: Mon Oct 22, 2018 9:53 am

Re: ZX Spectrum +2 booting problem

Post by CapSmasher »

Say no more, it works!

I just put a TIP31C in place of the ZTX653 (TR9) - ok I may overdo it with that radiator improvisation, but it was for testing :D
RAM seems to have those 4.7v - 4.72v but it seems it works like that. TEA2000 has around 10.5v or something

After 15 minutes these are the approximate temps (tested with IR temp with case open, ambient temperature of ~22):
CPU - 45 degrees
voltage regulator - 40
ULA ~38-40
TEA2000 -41
MC1488 - 42 (this one is hot compared with the other MC1489 which was pretty cold) - it may be this the problem that triggered the burn ZTX653 couldn't handle? Maybe if I use a device to test the connection? What device could I use ?
TIP31C - 25
RAM chips were pretty cold

Thank you for all the help and tips!

Image

Image

Image

Image

Image
CapSmasher
Drutt
Posts: 16
Joined: Mon Oct 22, 2018 9:53 am

Re: ZX Spectrum +2 booting problem

Post by CapSmasher »

I found out from the datasheet that the installed RAM (M3764) actually permits a minimum recommended voltage of 4.5 to 5.5 - and supports spikes of 7v (but for short time) maybe that explains why is still working. Maybe will help others this info if they will encounter the same problem.
CapSmasher
Drutt
Posts: 16
Joined: Mon Oct 22, 2018 9:53 am

Re: ZX Spectrum +2 booting problem

Post by CapSmasher »

Sorry to keep reviving this thread, but I have some new info that could help some inexperienced enthusiasts.
I found why it was only 4.7v on the RAM chips... Because I needed to verify first the voltage regulator, in all these tryouts I used some breadboard cables jumpers that seemed to drastically increase the resistance. After I soldered the original wires on the voltage regulator, 5v came back and all seems to function with normal parameters - the red led is not going on though - maybe it was burned or a loose wire (cassette player seems to work).

Can anyone give me some information about this 1488 chip? It seems that is heating as much as the CPU - pretty fast at around 40-45 degrees. The 1489 chip stays cold. It could be shorted? I have a lightgun, can this be used to test its functionality?
User avatar
1024MAK
Bugaboo
Posts: 3104
Joined: Wed Nov 15, 2017 2:52 pm
Location: Sunny Somerset in the U.K. in Europe

Re: ZX Spectrum +2 booting problem

Post by 1024MAK »

CapSmasher wrote: Mon Nov 12, 2018 8:23 am Sorry to keep reviving this thread, but I have some new info that could help some inexperienced enthusiasts.
I found why it was only 4.7v on the RAM chips... Because I needed to verify first the voltage regulator, in all these tryouts I used some breadboard cables jumpers that seemed to drastically increase the resistance. After I soldered the original wires on the voltage regulator, 5v came back and all seems to function with normal parameters - the red led is not going on though - maybe it was burned or a loose wire (cassette player seems to work).
Voltage drop via thin wires is not unusual. The circuit for the power on LED is very simple, just a resistor, the red LED and the wires/connections.
CapSmasher wrote: Mon Nov 12, 2018 8:23 amCan anyone give me some information about this 1488 chip? It seems that is heating as much as the CPU - pretty fast at around 40-45 degrees. The 1489 chip stays cold. It could be shorted? I have a lightgun, can this be used to test its functionality?
Keep in mind that one is a receiver only and the other is a transmitter/level shifter. So of course the temperatures they run at will be very different. For the 1488, can you test the input pins and the output pins with a multimeter on the voltage range. A short circuit on an output will cause the chip to get warmer than normal.

Mark
Last edited by 1024MAK on Mon Nov 12, 2018 2:02 pm, edited 1 time in total.
:!: Standby alert :!:
“There are four lights!”
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :dance
Looking forward to summer later in the year.
User avatar
1024MAK
Bugaboo
Posts: 3104
Joined: Wed Nov 15, 2017 2:52 pm
Location: Sunny Somerset in the U.K. in Europe

Re: ZX Spectrum +2 booting problem

Post by 1024MAK »

Links to three datasheets for 1488 chips:- One thing I forgot to say in my earlier post. 1488 chips are based on old bipolar technology. This type runs warm even when doing nothing. If you look at the datasheets, the typical power consumption/ dissipation is around 576mW. That’s over half a watt of power in a normal DIL case.

Mark
:!: Standby alert :!:
“There are four lights!”
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :dance
Looking forward to summer later in the year.
Post Reply