ZX Spectrum Issue 3 repair

[roko]

Hi Tezz,

In your message 8:19 you report you found 0 Volt at the upper right (RAM) pin, then find resistances at the other pins.
Do you realise that where semiconductors are involved all resistance measurings are indeterminate,
p.e. depending on the brand (unknown inners!) of chips?
(Pure shortcuts can be found, yes.)

Now you have removed the lower RAM chips you may concentrate on getting the voltage convertor working.
For checking that convertor you need to measure voltages anyway!
Maybe you could start measuring the output of (just) the transformer at its plug? Red probe at the inner contact, black at the outside.
Meter in the 20V= position (eventually in 200V= pos.). Measure again with probes exchanged, to see the '-' sign displayed.

Next step would be measuring the in- and output voltages of the 7805 regulator. (Leave ULA and RAMs out.)
You got a PM!

roko

[Tezz]

Hi Roko, thanks I've sent a reply to your PM.

I made sure to have the power supply that I'm using for testing fully serviced and recapped last month to be certain that it wouldn't cause any issue. With the 7805 regulator removed from this board now I have been testing the voltages and resistance again. The +5v line not giving the expected reading when the regulator was in place and also now it's removed is what I'm trying to understand at the moment. I was hoping that there would be a short in a failed lower RAM IC causing the issue. I'm not sure with the regulator removed and no lower RAM what the 5v line should now show on the meter

[1024MAK]

Hi Tezz

Apart from a multimeter, do you have any other test gear?
Specifically, do you have a current limited "bench" power supply. Ideally one with an adjustable current limit control?

With regard to your issue 3 board, I'm still concerned that "the lower right 5v pad on 20k range shows very low however 0.01". Does this still give a result of about 2.6 ohms on the 200 ohm range?

If yes, it is looking like the problem is most likely a solder splash or other short circuit between the +5V supply tracks and the 0V / ground (GND) tracks.

If you have an adjustable current limited "bench" power supply, one test would be to set the current limit to 700mA. Then connect the positive output to the Spectrums +5V power rail and the 0V (negative) to the Spectrums +0V rail. Only if the resistance between the +5V rail and the 0V rail is 2.6 ohms, this would result in the current flow trying to be 1.92 Amps! Which is rather more than the normal current for a Spectrum (typically about 500mA to 650mA). The 7805 voltage regulator is rated at 1A. And the mains adaptor / PSU is only rated at 1.4A.

Finding short circuits is not easy. The first step is to very carefully visually inspect the tracks, pads, component legs on both sides of the PCB. Unless you have young eyes, preferably with the aid of a good light and a magnifying glass.

Mark

[Tezz]

Hi Mark,

I've just retested the 5v with the meter set to 200ohm, it reads 07.5
I had a good long look at both sides of the board in the light and can't see any sign of stray solder or anything obviously untoward other than a couple of caps C25 and C65 looking suspect with their positive (axial) ends bulging. I was intending to change all the caps after the faults had been diagnosed and fixed but might they be an issue with the readings? Does the result of the 5v reading still look too low? I'll get the magnifying glass out later if there is still a suggestion of a bridge.

[1024MAK]

Yes, 7.5 ohms is still a rather low value.

Electrolytic capacitors typically fail with a reduced capacity or open circuit. But any that look distorted or are leaking need renewing anyway.

Ceramic capacitors that show visible signs of damage sometimes go short circuit.

Mark

[Tezz]

Thanks Mark, I'll look to see if I can find any breaks or stray solder with the magnifying glass in the morning. Not sure what to do next if I don't spot any problem there. Can ceramic caps be tested whilst in circuit? I forgot to answer regarding the bench supply earlier. I don't have one unfortunately. I should really look at buying one to undertake these repairs.

[1024MAK]

For digital boards where the majority of ceramic capacitors are used to "decouple" the power rails, no, there is no practical way to test them in circuit. For most types, failure is rare. So just have a good look. The disc type are most likely to suffer damage. The service manual recommends replacing them with the type that looks like resistors. But this is not vital.

Decoupling capacitors fitted across power rails, are there to supply short bursts of electrical energy to supply nearby chips when those chips need a short burst of power. For example, when a logic gate changes state. By having a capacitor near the chip, this reduces the peak load on the power supply network (PCB tracks, wiring and the power supply). It also helps to reduce the voltage drop through the PCB tracks. This helps to reduce noise on the power network. Noise that otherwise could cause chaos with a microcomputer system or logic system.

Small value ceramic supply the short energy bursts. Larger capacity electrolytic are slower, but act as 'bulk' energy stores for the longer duration energy peaks.

Mark

[Tezz]

Brilliant, thanks for that Mark. It's a clearer explanation of the caps purpose. The explanation years ago I was given was that they basically smooth out and control the voltages within parts of the circuit.

[1024MAK]

Attached are two circuits that can be used with 9 to 12V power adaptors and which offer a basic current limiting function. Note that these are both just "paper" designs, but should work okay (these exact circuits have not been tested, but I have built designs based on the chips used in the past).

The best version is the L200 circuit. This should work okay when fed by a Sinclair ZX Spectrum 9V 1.4A PSU if you want to use it with a 5V (or lower) output voltage. Higher output voltages require a mains adaptor/PSU with a greater output voltage. Remember to select the current limit and set the output voltage to the correct value BEFORE connecting any device or circuit to the output.

L200 based PSU.png (44.47 KiB) Viewed 3230 times

The LM317T version needs a 12V PSU as it needs a minimum input voltage of 11.25V. It may also work at low currents with a Sinclair ZX Spectrum 9V PSU, but cannot power a ZX Spectrum when fed by a Sinclair ZX Spectrum 9V 1.4A PSU.

LM317T current limiter.png

(79.64 KiB) Downloaded 171 times

All the regulator chips need to be fitted to suitable large metal heatsinks.

The LM317T should be available at any good electronics supplier. The L200 is slightly harder to get. Farnell do stock it, but I have not checked other suppliers.

Mark

[Tezz]

Thanks Mark, the L200 based design you have created looks like a good solution. Are you intending to put one together yourself or was it just an exersise in creating the circuit design? Would the 81 heat sink be sufficient? I should have a few of them spare after changing the boards to the DC converters.