ZX81 weird tape loading issues

[SiriusHardware]

Falter: If you mean something like a hi-fi cassette deck 'separate' then I'm afraid you'll have problems there as well.

The 'Tape out' or 'Line out' jacks on the rear of the deck will usually be at a fixed level, almost certainly not the level you need. (Occasionally, the output level is variable).

If the deck has a phones-out jack as they often did, it's very unlikely to be powerful enough for your needs. Only the phones - out socket on something designed to directly drive a speaker (internal or external) is likely to have enough power.

Thin black lines half-heartedly waving about in response to the input is a pretty typical symptom of the input signal not being at a high enough level.

You're right that if you are using a stereo source it's generally best to use only one channel output to drive the ZX81 because any phase differences (or any differences at all) between the right and left channels will mangle the signal that the ZX81 is expecting to see. If you only use one channel output or the other, then any differences between the two outputs will not cause problems.

You were going to post one of your .WAV files here, why not do that and then at least we can see if your file is OK and rule that out?

Cthulhu: The ZX81 was designed to work with only one thing, a typical contemporary mono cassette tape recorder / player. All the other things we're discussing here (Laptops, MP3 players) obviously didn't exist. Regarding the transformer, can you suggest a suitable part number from a USA / Canadian supplier? (Falter looks to be in Canada, I am in the UK but we don't know where you are).

 

[cthulhu]

The ZX81 was designed to work with only one thing, a typical contemporary mono cassette tape recorder / player. All the other things we're discussing here (Laptops, MP3 players) obviously didn't exist.

No, but the -10 dBV consumer line level standard certainly did. Virtually anything modern or vintage with a line or headphone out will provide that at least. Not hot enough for a ZX81 but why didn't Sinclair ensure it was?

 

[billdeg]

I'd like to re-state - you can still buy old style tape recorders with the appropriate jacks built into the side for cassette storage. Given they're cheap, I suggest this is always the best option. Using a vintage tape recorder is less reliable. New ones still look exactly the same, no one has to know if you're doing a public exhibit.

Maybe someone here can measure their tape recorder output voltage (after successful read) and share with group so as to advise what the target levels should be for whatever device you're using to drive a program input into the computer.

If you're trying to make backups of tapes, you don't want to send headphone driving voltages into a modern computer input jack, unless the volume is set to (practically) 0.

 

[SiriusHardware]

Not hot enough for a ZX81 but why didn't Sinclair ensure it was?

Because that would have needed more components and contravened Sinclair's prime directive, namely, to do as much as possible with as little as possible so that the item could sell for a set (low) price.

Everything Sinclair ever made was a minimalist masterpiece, often to the detriment of specification / performance / common sense, always with the final price to the consumer uppermost in mind.

 

[SiriusHardware]

I'd like to re-state - you can still buy old style tape recorders with the appropriate jacks built into the side for cassette storage.

Indeed, but the problem for Falter, as already mentioned, is that the programs he wishes to load exist only in the form of a digital audio file, which is why he's trying to play them out of a computer into a ZX81. He does not have any tapes, as I understand it, with ZX81 audio files on. Therefore, a tape recorder would not, in and of itself, be a direct solution to his problem.

You could (as you may be suggesting) play the audio file out from a computer into a traditional tape recorder to make the file into a tape, in which case it would be better to choose one (vintage or not) with a line-in input as well as a microphone input and use the line-in input to record material from the PC. As you rightly say, the output from any phones output on a computer or anything else will be much too loud to feed directly into a tape recorder microphone input which expects an input in the order of millivolts.

In the interests of science, I'll fire up one of my two ZX81s (neither are within immediate reach at the moment) and see if I can suggest a typical EAR input voltage to aim for. It'll take me a couple of days to pull it all together.

The easiest solution, if it works, and the theory is good, is to use Cthulhu's passive audio step-up transformer idea. The trouble is, those items are not just something you can easily buy from an electronics store nowadays.

Failing that, an active solution: Use a small audio power amplifier IC to lift the audio output level from the PC / MP3 player output to loudspeaker level audio. Here's a randomly found example:

http://hackaweek.com/hacks/?p=131

This amplifier could, possibly, run on the unregulated DC voltage from the ZX81's AC power adaptor (make a Y-splitter to send power both to this circuit and the ZX81) but I would play safe and run the amplifier on a battery for initial experiments at least.

 

[SiriusHardware]

Ok, a bit longer than a few days, sorry about that. The short answer is that you need a minimum of about 2V peak to peak coming out from your audio source into the 220 ohm load of the ZX81 EAR input. I found that once you've hit the minimum level you can take it quite a bit higher and it will still work, so it's achieving the minimum level which is the important thing.

I discovered I didn't have an LM386 amplifier chip lying around, went online to look for one and instead ended up buying a little LM386 based mono amplifier kit which included a PCB and a prewired audio input lead (and a speaker, not needed for this although handy for testing the amp). It took a while to arrive, hence the delay.

Initial evaluation of the Amp was done with it running on a 9V battery and with a variable output sinewave signal generator as the input signal. With a 220R resistor (same as the ZX81 input resistance) connected to the Amp output I slowly increased the drive from the generator until the point where the output waveform started to clip at the top and bottom edges - this happened at around 8V peak to peak output into 220 ohms.

So, with the addition of this amplifier I now had up to 8V of undistorted audio output to play with. I connected my netbook's audio-out to the input of the amp with a plug / cable wired to pick up only one of the two stereo output channels, the other not connected. With a 3.5mm mono jack socket on the output of the Amp, I used one half of the standard ZX81 tape audio lead to connect the output of the amp to the EAR input of the ZX81.

Next stop was this fantastic little site:

http://www.zx81stuff.org.uk

Here there is an amazing collection of ZX81 software. The site also offers a full speed online ZX81 emulator complete with a full onscreen ZX81 keyboard (or just use your computer's keyboard) so if you see a title you'd like to try out, it's as simple as clicking the 'Play' button next to the software item's entry. There's also a 'download' button next to every entry so you can get the software as a .P or.TZX format file.

Once you've picked something to download (I chose 'Seawolf', a nice 'Silent Service' style submarine game) - if you then go to 'Utilities' on the site you'll find a bit of software there called 'TapeUtils', which can either make original ZX81 audio recordings into .P or .TZX files, or, it can take a .P or .TZX file and either save it as an audio file or play it straight out of the computer's sound output.

TapeUtils is a .jar (Java) program so you may need to install the Java runtime environment if you don't have it already.

Having set all this up I loaded Seawolf.tzx into TapeUtils, set TapeUtils away playing the file and adjusted the computer's audio output volume until the output from the LM386 amplifier with the ZX81 connected was about 4V peak to peak on my scope.

Having done that, I stopped the playback process on TapeUtils, and put the ZX81 into load mode (LOAD""<newline>). The ZX81 started to show the usual 'no-signal' thin black diagonal lines.

I restarted the file playback and the display changed to show bold, constantly changing horizontal black lines... this went on for a few minutes, at the end of which.... Seawolf had loaded, just like that, first time.

After that initial success I tried it over and over again with different audio levels, and found that it would actually work with the Amp audio output level anywhere between roughly about a minimum of 2V peak to peak and the maximum available 8V peak to peak. Anywhere below 2V, though, and the signal became too weak for the ZX81 to read reliably - and bear in mind that it has to be 2V or more into a 220 ohm load (not into no load).

 

[cthulhu]

That was an interesting read - thanks for testing this. A 2 volt signal into a 220 ohm load requires a minimum of an 18 milliwatt (0.018 watts) power output from the playback device. This power output is within the capabilities of some portable media players but as I mentioned in a previous post a transformer suitable for impedance conversion will be needed. This is because most portable media players capable of outputting 18 milliwatts will only be able to do so into a low impedance load, e.g. 16 to 32 ohms typically. The correct transformer will make the 220 ohm load of the ZX81 look like this typical low impedance load to the player.

 

[cthulhu]

Ignore my previous post as I got the calculations wrong. I just re-read SiriusHardware's last post and he specified 2 volts peak to peak was the minimum signal level required. My calculations were for 2 volts RMS, 2 volts peak to peak is 0.707 volts RMS which means only 2.2 milliwatts is required from the media player. This power output should be well within the capabilities of every portable media player. A media player capable of outputting of 18 milliwatts into 32 ohms would not require the transformer at all as this is already more than 2 volts peak to peak.

 

[SiriusHardware]

I think I originally mentioned that I had tried this quite a while ago with an MP3 player (a Creative Zen Nano Plus, for information) with no luck. This player and others similar to this usually only run on one 1.5V AAA cell, although they may employ an internal DC-DC converter to maintain the unit's internal supply voltage at some other nominal fixed level as the cell voltage declines.

Not knowing what the internal supply voltage of the unit is, I can't say whether that particular unit is actually capable of outputting 2V peak-peak into any load, regardless of its nominal output power capability into its normal 32-ohm load. Incidentally in Europe / UK where I am, the maximum audio output level from personal in-ear audio players is legally restricted - but to exactly what level, I can't remember.

I'll use the 'make a WAV' feature of the TapeUtils software to create a .WAV of the same software and see what the maximum peak-peak voltage out from my Creative player is into 220R.

I do also have some, now not readily available 'Red' audio transformers of the type which typically used to be used in 1970s / early 1980s transistor radios with push-pull output stages.

I tried the active amplifier method first because that's something which everyone will be able to recreate now, either by building their own LM386 amplifier from scratch, or a kit similar to mine, but in the interests of scientific enquiry I will connect the output of one channel of the Creative player to the low impedance (speaker output) side of an audio transformer and measure the voltage available from what was originally the input / primary / medium impedance side of the transformer with a 220R resistor connected across the outer ends.

I'll report back when I have more.

 

[billdeg]

Being a musician, there is a pedal that I use to convert instrument (low) voltage into line-level voltages. It's called a SansAmp. Maybe you can find something like that, 9V, to boost and control the levels. Might be worth investigating.

 

[SiriusHardware]

Billdeg, what you describe is essentially what I did, used a 9V powered amplifier between the insufficiently powerful output of the player and the input of the ZX81. As the output from your SansAmp is only line-level, I'm not sure whether it will be able to deliver 2V or more peak to peak into a 220R load, but why not try it?

Further experimentation with the Creative player is currently curtailed because I mysteriously can not find it. I had it in my hands not two days ago.

Just tried to attach a .WAV file but find that the forum does not appear to consider .WAV a valid file format for attachments. I don't really want to convert it to a lossy format which may add distortion and, amazingly, I have nothing on this computer which will allow me to add it to a .zip file either. (I've only got the standard Windows XP zip extractor).

I'll be back when I've found the player and done some further experiments.

 

[SiriusHardware]

Having found the Creative Zen Nano plus again, I made some more measurements. With the .WAV file created by TapeUtils loaded, the maximum output from one channel to ground with a 220 ohm load was measured at 1.0V Peak to Peak: Not enough.

My player had the European firmware it was sold with and for some time now the maximum output volume on personal audio players in Europe has been capped for legal / safety reasons. I reflashed it with the USA version of the firmware, thankfully still available from Creative's USA website. This brought the maximum output into 220 ohms up to 1.3V peak to peak. Still not enough. You'll recall that I found earlier that the input to the ZX81 needs to be at least 2.0V peak to peak, possibly a little more.

So then, I took the original .WAV file and loaded it into 'Audacity', converted it to stereo with both channels identical, inverted / mirror imaged only one channel, and saved the file like that: Stereo, with the channels exactly in antiphase.

With the audio file modified this way, one channel output 'pushes' while the other one 'pulls', and vice versa. In other words, the two independent amplifiers are being used in a 'bridge' configuration, co-operating to produce a larger output swing than either of them can manage individually. A 220R resistor was connected between the left channel output and the right channel output: The voltage swing across it at maximum volume was from +1.3V to -1.3V, or 2.6V peak to peak. Comfortably enough to work, or so I hoped.

The Player-to-ZX81 lead was rewired to send the left channel output of the player to the 'signal' terminal of the ZX81 EAR socket, and the right channel output of the player to the GND terminal of the ZX81 EAR socket, and the 0V / GND terminal of the player's output socket was left disconnected.

I set the ZX81 into load mode and found by experimentation that the best looking 'load picture' was obtained with the volume at around 24-25 out of 25, with the output drive at that volume significantly more than the 2V minimum required.

However, there was one last obstacle to overcome.

Although the Creative Zen Nano Plus CAN play .WAV files as long as you place them in the 'Recorded Tracks' folder, for some reason, the ZX81 didn't like the way the Zen Nano Plus was playing back the .WAV files and rarely worked: The screen would keep flickering all the way through the load sequence, but as the audio file ended the computer would just reset.

The same antiphased .WAV file worked fine from my Dell netbook, either when I played one channel only through the LM386 amplifier as per earlier experiments, or if I dispensed with the amplifier, ran the netbook on battery power only and drove the ZX81 directly from the antiphased stereo headphone outputs as described above, so I knew the file was OK.

On a whim, I loaded the antiphased .WAV file back into Audacity and converted it to an MP3 and sent that to the Zen Nano Plus. This, to my complete surprise, worked every time with the volume set to 24 or 25 out of 25. I don't know what the ZNP does when playing WAV files but the ZX81 doesn't like it. I've captured the .WAV playback output on a storage scope and it looks really pure. But for whatever reason, the ZX81 likes the ZNP's playback output from the MP3 version better.

Anyway, to summarise: If the output from your player device is too low to drive the ZX81 directly:

-Make the mono .WAV file into a stereo version of itself with one channel a mirror image of the other, so that the two channels are in antiphase
-Wire audio source left channel out signal to ZX81 EAR input signal terminal
-Wire audio source right channel out signal to ZX81 EAR input ground terminal
-Do not connect the audio source's audio output GND to anything
-Ensure the audio source (laptop, MP3 player) is running on standalone battery power and disconnect it from power supplies or any other devices which might connect the 0V of the source to earth.
-This is necessary because although the ZX81's own PSU is an isolated transformer type, the ZX81 may still find a path to ground through the outer coaxial connection of the aerial / antenna socket, depending on the TV being used.
-If you're getting the required 2V+ Peak to Peak into the ZX81 but it still isn't working, try converting the antiphased .WAV file to MP3.

If you're using a computer rather than a dedicated audio player for this, there's one more thing to watch out for: Any extraneous sounds produced by the system, like the gentle 'bong' you get when an email has just arrived in your inbox, will break the loading process if they happen to occur while you are trying to play a file into the ZX81. To avoid this annoyance you'll need to go around disabling all the system audio alerts so this doesn't happen - much easier (and neater) just to use a dedicated MP3 player.