A500 Parallel port has bad strobe/pulse

[Jake001]

Hello all!


I am trying to built an audio recorder expansion for my Amiga 500 and the audio it recorded was very distorted and awful. I scoped the strobe pin on the parallel port and it is all over the place. It'll hold a 5v pulse for a few seconds then drop to around 0.5v then go crazy.


Does anyone know how the Amiga generates that pulse and how to fix it??


Thanks!
Jake.

[Toni Wilen]

CIA-A chip generates it automatically internally when reading or writing parallel port data. Swap CIA chips (CIA-A in place of CIA-B and vice versa) and check if problem goes away or changes.

[Jake001]

Does this thing affect anything?
Its positive (red) lead is attached to pin 8 on CIA A.
Looks like some sort of ROM expansion.

[Jake001]

I haven't swapped the chips yet as I'm still trying to find a good way of getting them out without damaging the pins.

[Jope]

Small flat bladed screwdriver between the chip and socket, lever up 1mm at a time from alternate ends. Make sure the screwdriver tip does not scrape the PCB under the chip.

[Vypr]

That board looks like a ROM switcher tied into the CIA chips to allow ROM switching by keyboard reset

[Jope]

Pin 8 is Fire0, so hold down the left mouse button while reseting to get a rom switch.

[ElectroBlaster]

Not really relevant to the problem, but I noticed the rom switch on the second picture, one of the kickstarts has written on it:

"CREDIT CARD INTERFACE 2.05"

Whats the score with that? I gather its a standard 2.05 kick rom, but never seen one with writing on it like that before...

EDIT: This rom corrected the early A600 kickstart/ide/pcmcia slot bug?

[Jope]

It's the 37.300 ROM, which was the first to include PCMCIA support. Most of the chips don't have that writing, it could be that the fab that made this particular chip included too much text from the accompanying documentation.

[Stedy]

Originally Posted by Jake001

Hello all!


I am trying to built an audio recorder expansion for my Amiga 500 and the audio it recorded was very distorted and awful. I scoped the strobe pin on the parallel port and it is all over the place. It'll hold a 5v pulse for a few seconds then drop to around 0.5v then go crazy.


Does anyone know how the Amiga generates that pulse and how to fix it??


Thanks!
Jake.

How did you measure the /STROBE signal and did you have a nice short ground lead or the typical 75-120mm ground lead?

With a fast edge and a bad grounding arrangement, the signal will look utter rubbish if incorrectly probed.

Might be worth injecting a sinewave into the audio circuit and see if it gets distorted by the input circuit, you may have an issue in the analogue domain.
The Amiga parallel port is current limited, 10-20mA springs to mind, if you draw too much, the supply will drop, causing weird effects.

Can you provide the circuit you are using?

[Mad-Matt]

Originally Posted by Jope

It's the 37.300 ROM, which was the first to include PCMCIA support. Most of the chips don't have that writing, it could be that the fab that made this particular chip included too much text from the accompanying documentation.

37.299 Was the first with pcmcia support but for some reason missing scsi.device for the ide. I had to Boot from a floppy which included scsi.device from kickstart 3.0 which mounted the hd and transfered control to it and booted wb as normal. pcmcia was used for squirrel cdrom.

[Jope]

Originally Posted by Mad-Matt

37.299 Was the first with pcmcia support but for some reason missing scsi.device for the ide. I had to Boot from a floppy which included scsi.device from kickstart 3.0 which mounted the hd and transfered control to it and booted wb as normal. pcmcia was used for squirrel cdrom.

Indeed, you are right. I remembered wrong.

[Jake001]

Here is the circuit for the audio sampler.
I am using a LM358 Op amp to scale the input signal between 0-5 volts and a Max153 8-bit ADC. A general purpose hex inverter inverts the strobe before it is sent to the ADC.

[Jake001]

[Jake001]

Originally Posted by Stedy

How did you measure the /STROBE signal and did you have a nice short ground lead or the typical 75-120mm ground lead?

With a fast edge and a bad grounding arrangement, the signal will look utter rubbish if incorrectly probed.

Might be worth injecting a sinewave into the audio circuit and see if it gets distorted by the input circuit, you may have an issue in the analogue domain.
The Amiga parallel port is current limited, 10-20mA springs to mind, if you draw too much, the supply will drop, causing weird effects.

Can you provide the circuit you are using?

I put different frequency sine waves through it and you can hear them but there's a lot of distortion. If I measure the input audio with my scope it looks fine so I think the ground is sufficient.

[Jake001]

This is a little drawing I made of what the sound input signal looks like on the oscilloscope on the Amiga itself while it's recording. Some of the signal is recorded fine but a lot of the data is flat at 0. To be clear the signal going into the Max153 from the op amp is clean.

[Jake001]

I put a 10k resistor in line with the strobe to limit current but that doesn't seem to affect anything. [ Show youtube player ] here's a video of what the oscilloscope on the Amiga looks like. It seems to change based on the input signal which is weird. (I was just touching the end of the cable to make some random noise) It kinda looks like the ADC chip (Max153) is only sending data when there's a change in the input, but that shouldn't be the case.

[Jake001]

Originally Posted by ElectroBlaster

Not really relevant to the problem, but I noticed the rom switch on the second picture, one of the kickstarts has written on it:

"CREDIT CARD INTERFACE 2.05"

Whats the score with that? I gather its a standard 2.05 kick rom, but never seen one with writing on it like that before...

EDIT: This rom corrected the early A600 kickstart/ide/pcmcia slot bug?

So I held the left mouse button on reboot and it did indeed switch the ROM. What this does is allows to either boot workbench from the hard drive I have installed or from a workbench disk.

[keithamiga1200]

A pair of long tweezers are good for lifting a chip.


That's an interesting sounding project (excuse the pun)

[Stedy]

Hi Jake,

What I'm about to post below, will seem a little bit like a teachers review of your design, please take it in the spirit of I saw a few things and I corrected them. I've been doing Analogue electronics for 14 years now, learnt a few things.

Attached is a marked up copy of your circuit diagram with some notes. I'll explain them here.

Note #1 Smoothing capacitor in the wrong place. You need to remove it otherwise the reactance of the capacitor will change the gain of the amplifier, dependent on the frequency.

Note #2 Your amplifier has a gain of 2. For a non-inverting amplifier, the gain is 1+Rf/Rg where in this instance, Rf=Rg=22K. The LM358 can only handle 0-4V on it's input and output 0 to 3.5V when running on +5V only. This is hidden in the datasheet. It's nearly 40 years old. Remove the lower 22K resistor for unit gain if needed.

You also need a load resistor to ground, hence the 10K added. I had a similar issue 6 years ago https://ianstedman.wordpress.com/201...e-simulations/

I use the MCP602 amplifier, it works better than the LM358 in single supplies, has a 0 to 3.8V input and 0-4.90V output range

Note #3, you have the wrong mode selected on the MAX153. With pin 7 (Mode) floating, it defaults to mode 0. This causes it to drive out on pin 6 (RDY), which will conflict with the /STROBE signal from the Amiga. Pull pin 7 to +5V and pin 8 (RD) to 0V then you can use just pin 6(WR), to initiate the conversion when /STROBE goes low for about 1uS and 250ns after /STROBE goes high, you can read the parallel port inputs to get the 8 bit value.

Note#4, you don't need the inverter on /STROBE. On the falling edge of strobe, with my revised circuit, you instruct the ADC to sample. When strobe goes high, it will output after 660ns, the converted data. With your setup it's always trying to convert but not always outputing data.

Note #5, add more capacitance. This will help smooth out supply variations and give cleaner results. Move the 2.2uF capacitor you took out in note #1 to be across the +5V supply. Also add another by Vref (pins 11& 12) to smooth out noise from the Amiga.

That's todays lesson over.