128k +2 (grey) Ear mod - New ULA and still problems
Bverstee kindly sold me a 128K ula since the one in my +2 seemed to have a broken ear in. But even with the new ULA, I still can not load games through my added EAR input. I dont have any ZX tapes so I can not even test if it can load games from the datacorder properly. I have borrowed an oscilloscope and I measure the input on Pin 34, what reaches it is around 500 mV. I have checked the resistors etc. against the schematic, and they are ok.
Input on the white cord (EAR from casette) is around 3.6V signal.
Any suggestions?
Input on the white cord (EAR from casette) is around 3.6V signal.
Any suggestions?
Post edited by Johan1973 on
Comments
I think the output dac is not good enough and the sound is not dynamic enough.
First thing I would do is to obtain a tape and save something on it with the built-in-datacorder.
Then load it it back.
After that you can wire the connector from the datacorder to your ear mod.
Then load what you saved.
If you have that working, you know for sure that the circuits are ok.
The built in datacorder is toast. Something extreme has happened to this +2 in the past, because much of the circuits have been badly replaced in the past, and I find that the LA6324 in the Datacorder is fried. So I simply removed the datacorder for now.
If you can save something from the grey+2 to your pc, can you then load it back in one of the other spectrums ?
Or even better, load it back ?
Again to try and see what might be going wrong.
The sound card I am using is a Soundblaster X-FI PCI-E which I have completely overhauled and put decent DACs on for both playback and recording of audio.
Even that said, the meager built in realtek audio of my motherboard is sufficient to load software from, though I have never tried saving anything to it. And that is quite poor quality sound hardware.
The level and clarity of the signal is not the problem here. Checking input signal with oscilloscope I get a perfect clear signal through my added EAR port. The problem is when it reaches C49 just before ULA. Before it the level is a decent 3.6+ V level, but after it and on PIN 34, it is around 500mV.
Just yesterday I overhauled the board and replaced all electrolyte caps on the board. So I know it is not a bad cap (C49 is one of those I replaced).
On my +2 I connected the new added EAR-MIC socket to the pins of the datacorder header.
Shouldn't the EAR line go via R47 then ?
EAR socket is connected to the white wire that is input from data corder. Since the LA6324 amplifier is toast. I have tried using a buffer transistor to get the signal to a higher level. It produce a higher level signal, but still, there is no sign that the ULA receive signal other than at around 500mV level. And it is not loading anything.
But another question: Is R102/390Ohm placed on the solder side of the board on your +2, or is it placed on the component side as R82?
If it is missing it could be the cause.
I guess I need to solve on problem at a time, and then test the old ula again.
R47 input side: 6 Volt signal
R47 ULA side: 2 volt signal
If I go any lower, it won't load.
Used Flying Shark game and it has a custom loader.
Thank you, that helps a lot to know what it should be at when working. I will do further testing. I put the original ULA back in the machine now since the replacement was not working, screen was flickering, not syncing, and random static "noise" appearing in picture.
If the signal at the ULA becomes too much of a sinewave instead of a square wave, loading fails.
And this happened on my grey+2 when it got lower than 2 volt.
*deleted due to fault*
Do you have value for R5 ?
I think I have the components and I'd like to try the circuit with my el-cheapo MP4 player to see if I can get it to load something.
Sorry, I forgot to put the value of R5, R5 is 470 Ohm.
UPDATE: I have fixed the problem in the schematic and given R5 a value. the BC557 was turned the wrong way, emitter and collector had switched place on how they where supposed to be connected.
UPDATE2: For anyone viewing this and not having seen the original deleted schematic, I add; V+ is input from EAR input port, and V- is output to Pin 4 on cassette input connector that goes to ULA.
As far as I can see at first glance, Q1 (just) creates pulses unless input V+ apart from sound also holds a DC voltage.
My calculation goes somewhat like this: Vb=(470/940)*5=2.5V, so Ve=1.9V, so Ie=Ic=1.9V/1.5K=1.3mA, so V_R1=1.3*4.7=6.1V which requirement cannot be fullfilled.
Another IMHO problem is that the collector of BC557 (=PNP) should have a negative voltage compared with emitter in order to conduct (and work).
These remarks are not ment negative Johan, just my puzzling observation after I noted the quite low value of R3,R4.
I am just a beginner at electronics, but I have used this design before. The idea behind it is that Q1 which is biased create a constant low current over R1 and R2. R3 and R4 biases Q1 so it reacts much more quickly to any change in the input voltage of the signal.
Q2 which is a PNP is controlled by the drop in voltage from VCC over R1 (inversion). Q2 then sources VCC to create an output pulse that is relative to the input pulse, but is lifted to a higher voltage.
I have not done any math on it. I have just tried and measured levels to find what produced the level I required.
This particular circuit I have used before in touch activated controls for a modified xbox 360 controller, with great results. I am 100% certain it can be optimized if I had a bit more knowledge about electronics and did the calculations.
If you have suggestions and knowledge to share, I am happy to listen and learn :)
the first schematic I understand a bit. This one is faulty.
If Q2 is really a BC557 it is a pnp type. In that case the emitter has to be more positive than the collector and the base.
Furthermore you should think about adding a capacitor in series to the input. Your input line is held at about +2.5V. If you the output of the sound source
is sensitive to DC voltage damage would be possible. Also a resistor to ground at the output of your sound source would shift the base voltage of your Q1. Based on your resistors Q1 is saturated normally.
I would set the voltage at the base to a lower voltage than 2.5V.
Here is a suggestion based on your parts:
I just had a similar problem while building my harlequin.
I hope not to offend you - forget it if you are satisfied with your solution.
Greets Ingo.
I have treated the transistors as "switches" mostly.
And if I understand your circuit correctly it is a good design, since each transistor handles one part of the curve (positive and negative) while balancing them and lifting voltage on output based on R3 and R4.
I only need a 7805 regulator and a 9 volt battery now.
Tomorrow I'll test the circuit.
Output on my mp4 player is under 5 volt max, so I'll see what happens.
I'll also check with the BC557 reversed.
And I will try the circuit that Ingo suggested.
All of this gave me an idea though.
If one would use a logic gate, protected via a resistor or voltage divider and a zener diode at the input.
Then, at the output use something with an open collector or an open emitter.
Will that work ?
UPDATE: Fixed error (schematic was not as I had wired it on breadboard)
Need to transfer this to PCB now. I hope someone can give me some feedback on it, to make sure I am not screwing up royally..
When you want to amplify 'ups' and 'downs', then (here), without incoming sound, the voltage at the collector should be somewhere halfway the minimum and maximum, 0 an 5V here. So that there is room for the voltage to go up and down. When Vc, V(oltage)(at)c(ollector), without signal is found close to 0V or 5V then there is no room left for 'downs' (0V) or for 'ups' (5V). Sometimes this is what you want, the transistor acts like a switch then. A specific working point is chosen by biassing the basis. Re, R(esistor)(at)e(emittor), is there to stabilize the chosen working point. Ingo's diagram, where T2 'feeds' T1 over R1, brings a much better stabilisation against changes in qualities (a.o. by heat) of the transistors.
More practical: I think that your diagram differs from reality as V- and V+ (the input and output) are connected to each other by means of (large) C1. Should V- and V+ really be the in- and output of an amplifier, then you would have a closed loop and the whole would be oscillating like mad, probably somewhere high up in the audio range.
But in the diagram V+ is not connected to any point that could act as output.
Should this diagram really depict the situation, then you might try again by using C1 only! Any way, coming to think of it, in cases where a 'tape' signal goes 'dead', one might want to do a test with large capacitors in both the signal and the ground wire. Just for preventing any DC levels playing a role!
I believe that long time ago (>10 yr) there was an 'amplifier' for headphones published in Elektuur's summeredition, which used inverting logic buffers.
I agree, when it works leave it.
That said, I will continue experimenting on my breadboard, simply because I want to learn more. I am however limited in what components I have, which is why my circuit have to work with what is available.