UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Specific Vintage Equipment > Vintage Computers

Notices

Vintage Computers Any vintage computer systems, calculators, video games etc., but with an emphasis on 1980s and earlier equipment.

Closed Thread
 
Thread Tools
Old 14th May 2023, 7:52 am   #81
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Micro music box project.

Well, since you have so many of those 16 ohm mini speakers how about chaining 4 to 6 of them in series to make a high impedance speaker? You'll need to wire each one with its minus terminal going to the plus terminal of the next so they all move in the same direction at the same time. Make this the collector load for an NPN transistor with its base driven via a resistor from the Flag output, and don't forget to put a back EMF suppression diode between +supply and the transistor collector, just as you normally would when driving a relay.
SiriusHardware is online now  
Old 15th May 2023, 2:22 pm   #82
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Micro music box project.

ICs have arrived, but my scope seems to have died, as has the battery in my Fluke meter.

Something tells me this is not the day to turn on my MK14...
SiriusHardware is online now  
Old 23rd May 2023, 1:15 pm   #83
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Micro music box project.

Sorry to have been quiet on this - have just powered up my PRBS generator at work and it isn't working - various things were leading me to suspect that the 4006 shift register might be a fake but then I went looking for the datasheet for the specific IC that I ordered - a Motorola branded MC4006P - and that turns out to be a 4-line to 8-line decoder. D'oh. Just thought I had better mention it in case anyone else stumbles down the same blind alley.

I'll try again when I actually have the correct (shift register) IC to hand. This time, I'll order from Cricklewood as I originally suggested.
SiriusHardware is online now  
Old 23rd May 2023, 2:54 pm   #84
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
Default Re: Micro music box project.

You could try connecting three of 74x273 as a replacement, but the 4006 will probably arrive before you finished wiring it.
Mark1960 is offline  
Old 23rd May 2023, 3:43 pm   #85
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Micro music box project.

Wiring up the current version was enough of a chore If I had realised that the one on Slothie-Ortonview was also based on the 4006 / 4070 I would have just moved the VDU switch to OFF (or temporarily removed the PIC) and I would have been there a bit faster.

But as it is, I still need a more 'accurate' 4006 IC anyway.
SiriusHardware is online now  
Old 23rd May 2023, 4:22 pm   #86
Slothie
Octode
 
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
Default Re: Micro music box project.

I have been slacking on this one, I've had a lot of back pain and its stopped me getting on with it which is frustrating because I think I all the elements in place now to try it out I just need to get things together! Hopefully I will get my act together soon
Slothie is offline  
Old 23rd May 2023, 4:52 pm   #87
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Micro music box project.

This could be the slowest race in history, as I also have a lot of things getting in the way at the moment. When I get to it I'll probably try your MK4teened version of the Twonky code first, as I have more confidence in your competence when it comes to that sort of thing.

I have, however, just ordered what will hopefully prove to be the right sort of 4006 chip, so I should be catching up with you soon.
SiriusHardware is online now  
Old 24th May 2023, 9:44 am   #88
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,392
Default Re: Micro music box project.

Quote:
Originally Posted by SiriusHardware View Post
Sorry to have been quiet on this - have just powered up my PRBS generator at work and it isn't working - various things were leading me to suspect that the 4006 shift register might be a fake but then I went looking for the datasheet for the specific IC that I ordered - a Motorola branded MC4006P - and that turns out to be a 4-line to 8-line decoder. D'oh. Just thought I had better mention it in case anyone else stumbles down the same blind alley.

I'll try again when I actually have the correct (shift register) IC to hand. This time, I'll order from Cricklewood as I originally suggested.
Quote:
Originally Posted by SiriusHardware View Post
>>
>>
But as it is, I still need a more 'accurate' 4006 IC anyway.
Quote:
Originally Posted by SiriusHardware View Post
>>
I have, however, just ordered what will hopefully prove to be the right sort of 4006 chip, so I should be catching up with you soon.
Yes, Motorola always put an extra '1' after their 'MC' prefix on 4000-series etc. CMOS logic IC's. So you actually need an MC14006 Motorola (Now mostly onsemi) part: https://www.alldatasheet.com/view.js...14006&sField=2
This often causes confusion in the opposite way, with people not realising the generic 4000-series number is with the extra '1' removed.

Surprisingly, I managed to find a quantity of the 4006B (as well as the 4030B Quad Ex-OR gate - marked as such - which I discovered that the 4070B is actually identical, as only the obsoleted original non-B CMOS ones were different to distinguish the older 4030 from the later 'issues-fixed' 4070 replacement. But they made these under each, for less-confusion over 4030 replacement)
The TWONKY parts list / schematic doesn't show the B suffix on these, but I assume they did use these (even if specifying the older 4030 number, rather than later 4070, might lead you to think old non-'B' versions were required).
If you do get stuck for a proper 4006B shift-register, I could probably post one to you.

So I thought I'd too have a go at building this two-IC PRBS generator, to see what the output was like. But then also found how bad the original article's 'circuit' diagram was - More of a block diagram, with lack of pin numbers / labelling. And you actually need to connect the 4006's four separate 4+1 bit shift registers together, externally, as they are not all connected-together internally as the TWONKY diagram seems to indicate.

I couldn't find any details of a 3-tap 18-stage PRBS generator online - only minimal (for a maximal sequence) two-tap tables of taps / Polynomials (Where there's a micture of starting numbering at 0 or 1, so often '1' difference between the taps / Polynomial numbers some give, relative to others) But I did discover that it seems only even-number of taps give a maximal-length sequence. And that the number of possible taps configurations increases if you do use more than minimal two taps for an 18-stage etc. This does lead me to believe that this doesn't use a maximal (for 18-stages) length sequence, and so not as random as it could be.

So that Elektor circuit that had been posted here was much-more useful in finding-out what was originally used, as it does seem extremely similar (except it used a -15V supply and a transistor inverter to make an Ex-OR into an Ex-NOR - rather than using a single +5V supply and a spare Ex-OR gate made into an inverter by pulling 2nd input high).
I wonder where the Elektor version originated, as it seems the more-standard minimal-tap maximal-length sequence details were in the CMOS cookbook / possibly the book on Linear Feedback Shift Registers (LFSR's).

I decided to place IC's side-by-side - like on Graham's layout, rather than one above the other, to make a smaller / neater layout on veroboard.
And by swapping the pin numbers over on the Ex-OR gate inputs, I found that minimised wire links between the two IC's with tracks doing most of the connections required. Luckily I sketched it out / optimised it on paper first, as I found a number of diagonal wire links were best placed under the IC sockets, before inserting these.
So I connected 'IC1' 4030 pins as follows:
p1 to 4006 p9+p5 (Tap at '5')
p2 to 4006 p13 (end of chain output)
p3 to 4030 p12
p4 to 4006 p6 (start of chain input) / Sense-B output.
p5 to 4006 p10+p4 (Tap at '9')
p6 to 4030 p11
p7 = Gnd
p&p9 (unused inputs) to Gnd
p10 n/c (unused gate output)
p11 to 4030 p6
p12 to 4030 p3
p13 to p14 (+5V supply)


Although I haven't yet got a completed MK14 (replica) and would need original / MK14 code modifying, due to memory-map differences, to get it to run on the PICL (NIBL) V2, under Kitbug+ that I now have one-of (even with adding an external oscillator circuit to the clock-input (as Elektor circuit has)) as the NADS signal is not available. But use of a constant-frequency clock rather than NADS may also affect actual-operation / resulting sounds.

So for now, I did manage to power-up this TWONKY-PRBS part only, that I'd built, last-week (see attached photos) and check it did something sensible..

I originally tried feeding a 1MHz square-wave in as the clock, as was only operating it off +5V, rather than 15V etc. for full-speed. And whilst it seemed to be working, with an approx. balance of Highs & Lows, I did notice some narrow (<1us) pulses with very-slow riise/fall times at the edges of these narrowest 'pulse'. So I then reduced clock to 500kHz (being as only operating from 5V where 4000-series is much slower), and I found the slow edges narrow pulses were still there.
This caused me to realise the narrow pulse was actually a glitch in this logic (possibly caused-by extra tap / use of extra inverter to make it a actually mixture of Ex-OR & Ex-NOR feedbacks - possibly a 'mod' to the circuit, to ensure this doesn't get stuck in an all-zeroes condition at start-up.

It seems either Ex-OR's or Ex-NOR's can be used, with the same taps, but does result in a different actual sequence (although the same total sequence-length). Ex-OR's / NOR's can also be used inline, between the LFSR stages, instead of just tapping and combining to feedback to single LFSR chain's input.
But I've not seen any info about using a mixture of types. And I suspect the odd-number of taps makes the resulting sequence rather-less than maximal for the 18-stages. To check this, would need to look for a repeat of an 18-bit state - which is difficult to do when you've only got access to every 4 bits in most cases. So might have to do this via some digital-logical simulation / doing it with a software equivalent, to dump to an entire sequence output file (I found these were available to download for various ones, on one webpage about LFSR's and PRBS generation).
Attached Thumbnails
Click image for larger version

Name:	IMG_8270__TWONKY-PRBS_on_Veroboard_(Topside).jpg
Views:	27
Size:	54.9 KB
ID:	278669   Click image for larger version

Name:	IMG_8271__TWONKY-PRBS_on_Veroboard_(Underside).jpg
Views:	26
Size:	66.6 KB
ID:	278670  

Last edited by ortek_service; 24th May 2023 at 10:03 am.
ortek_service is offline  
Old 24th May 2023, 10:33 am   #89
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Micro music box project.

Thanks for all the input Owen -

Quote:
Yes, Motorola always put an extra '1' after their 'MC' prefix on 4000-series etc. CMOS logic IC's. So you actually need an MC14006 Motorola (Now mostly onsemi) part
I think that somewhere in the dim haze at the back of my mind, I already knew that but I did what I should never really do, ordered them in a bit of a hurry late at night. Lesson learned. (For now).
SiriusHardware is online now  
Old 24th May 2023, 11:57 am   #90
Slothie
Octode
 
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
Default Re: Micro music box project.

Quote:
Originally Posted by ortek_service View Post

I couldn't find any details of a 3-tap 18-stage PRBS generator online - only minimal (for a maximal sequence) two-tap tables of taps / Polynomials (Where there's a micture of starting numbering at 0 or 1, so often '1' difference between the taps / Polynomial numbers some give, relative to others) But I did discover that it seems only even-number of taps give a maximal-length sequence. And that the number of possible taps configurations increases if you do use more than minimal two taps for an 18-stage etc. This does lead me to believe that this doesn't use a maximal (for 18-stages) length sequence, and so not as random as it could be.
The prng on Ortonview works - is that a 3 tap prbs?
Attached Thumbnails
Click image for larger version

Name:	prbs.jpg
Views:	30
Size:	57.3 KB
ID:	278684  
Slothie is offline  
Old 24th May 2023, 12:32 pm   #91
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Micro music box project.

The PRBS circuit in Slothie-Ortonview (post #90) is actually exactly the same as the one I linked to in #56 with the same interconnections, although the allocation of the individual gates from the 4070 may differ. That one seems to have been credited to Maplin, although it's impossible to know whether that came from a project, a datasheet or what.

I built mine from the Maplin diagram so it is coincidentally identical to the one on Slothie-Ortonview.
SiriusHardware is online now  
Old 24th May 2023, 12:47 pm   #92
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Micro music box project.

I forgot to say to Slothie: A good test for the crystal earpiece would be 'Music Box' although if you are running at 4.00MHz the rendition of 'God Save The King' may well sound more like the Sirius Cybernetics Corp's robot choir singing 'Go Stick Your Head In A Pig'. As far as I remember, the software timing for the note lengths and pitches in 'Music Box' is for the original 4.43MHz MK14 clock.

That might be a good little rainy Sunday afternoon job for someone, rework the timing and note values in 'Music Box' for a 4.00MHz clock.
SiriusHardware is online now  
Old 24th May 2023, 1:37 pm   #93
Slothie
Octode
 
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
Default Re: Micro music box project.

Quote:
Originally Posted by SiriusHardware View Post
That might be a good little rainy Sunday afternoon job for someone, rework the timing and note values in 'Music Box' for a 4.00MHz clock.
I'm currently using a 4.43 Mhz clock as I dont have a replica or original vdu that requires a 4Mhz clock. But the prospect of doing "Share and Enjoy!" authentically is attractive.
Slothie is offline  
Old 24th May 2023, 3:11 pm   #94
Silvester
Triode
 
Join Date: Feb 2014
Location: Redditch, Worcestershire
Posts: 27
Default Re: Micro music box project.

The circuit in #90 looks same as one used as noise generator in Maplin 5600S synthesiser (also ETI 4600), AFAIK originally designed by Trevor Marshall for ETI.
Attached Thumbnails
Click image for larger version

Name:	noise.jpg
Views:	28
Size:	66.9 KB
ID:	278700  
__________________
David
Silvester is offline  
Old 24th May 2023, 3:46 pm   #95
Slothie
Octode
 
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
Default Re: Micro music box project.

Quote:
Originally Posted by Silvester View Post
The circuit in #90 looks same as one used as noise generator in Maplin 5600S synthesiser (also ETI 4600), AFAIK originally designed by Trevor Marshall for ETI.
I seem to recall it was from the elector formant synthesiser which may have been bottowed from the 5600?
Slothie is offline  
Old 24th May 2023, 7:10 pm   #96
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
Default Re: Micro music box project.

Quote:
Originally Posted by Slothie View Post
Quote:
Originally Posted by ortek_service View Post

I couldn't find any details of a 3-tap 18-stage PRBS generator online - only minimal (for a maximal sequence) two-tap tables of taps / Polynomials (Where there's a micture of starting numbering at 0 or 1, so often '1' difference between the taps / Polynomial numbers some give, relative to others) But I did discover that it seems only even-number of taps give a maximal-length sequence. And that the number of possible taps configurations increases if you do use more than minimal two taps for an 18-stage etc. This does lead me to believe that this doesn't use a maximal (for 18-stages) length sequence, and so not as random as it could be.
The prng on Ortonview works - is that a 3 tap prbs?
18 stage with three taps might not be an optimum random prbs generator, and was probably just to use minimum number of chips, but I think the intention here is to replicate Twonky as close as possible to hear how good or bad a tune it can compose.
Mark1960 is offline  
Old 25th May 2023, 9:24 am   #97
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,392
Default Re: Micro music box project.

Quote:
Originally Posted by Mark1960 View Post
Quote:
Originally Posted by Slothie View Post
Quote:
Originally Posted by ortek_service View Post

I couldn't find any details of a 3-tap 18-stage PRBS generator online - only minimal (for a maximal sequence) two-tap tables of taps / Polynomials (Where there's a micture of starting numbering at 0 or 1, so often '1' difference between the taps / Polynomial numbers some give, relative to others) But I did discover that it seems only even-number of taps give a maximal-length sequence. And that the number of possible taps configurations increases if you do use more than minimal two taps for an 18-stage etc. This does lead me to believe that this doesn't use a maximal (for 18-stages) length sequence, and so not as random as it could be.
The prng on Ortonview works - is that a 3 tap prbs?
18 stage with three taps might not be an optimum random prbs generator, and was probably just to use minimum number of chips, but I think the intention here is to replicate Twonky as close as possible to hear how good or bad a tune it can compose.
Yes, I originally just wanted to build the TWONKY's PRBS circuit that was detailed here - Being as I already had the parts to hand, mainly to see how well (or bad!) it worked. Although I've just noticed there can be issues between using different CD4006 / HEF4006 & MC14006's makes and slope of input / might benefit from using an extra buffer gate : https://gr33nonline.wordpress.com/2018/08/19/lfsr/
(And, as well as mentioning why Motorola had to insert the extra '1' after 'MC' prefix, they also try to find actual differences between the 4030 & 4070 - So I might have to try swapping to a 4070, to see if that does actually make a difference in practice).

But due to how bad the original published schematic etc. was, did a bit of research into LFSR's, as although I was familiar with the basic concept (and were probably covered on my Uni's course, 30+ years ago, but I've no doubt forgotten a lot of that), I'd never actually built one myself.

I'd originally thought there would be just one standard implementation, but have now found-out there appears to be 7,776 different 'Primitive Polynomials' (= no. of taps configurations) that give a maximal length sequence on an 18 bit (degree) LFSR! - Although apparently maybe only 20 different configurations using 4-taps, that give a maximal-length sequence. And possibly only one configuration of the minimal 2-tap XOR's / XNOR scheme, for an 18bit LFSR.

So I was puzzled why they actually made the circuit more-complicated, with an extra-tap / extra X-OR gates, whilst getting a shorter length sequence than the maximal (2^18) -1 that is achievable with an 18bit LFSR.
It also seems using extra XOR's causes delays issues, restricting max. frequency with these ‘Fibonacci’ types. And so the 'Galois' type, with a single XOR gate between the stages is sometimes preferred (but might be a problem, if required points aren't available externally to the Shift register IC (the 4006 does mostly only have access every 4(+1) bit, but that can also be an issue with ‘Fibonacci’ type taps).

It might be that the extra complications are to try and ensure an all-zeroes start-up in the shift-register, doesn't get stuck in a continuous 0's which can happen on conventional maximal-length sequences, that exclude all-zeroes in them. But I'm not sure what happen's with this one if it starts-up with a code that isn't in it's non-maximal PRBS - It might cycle forever with the same code, or maybe there can be multiple much-shorter than maximal length sub sequences with this one? These do give some examples of what happens, with some non-maximal sequence designs:
https://sus.ziti.uni-heidelberg.de/L...9_DST/LFSR.pdf
https://case.edu/artsci/math/singer/publish/LFSR.pdf
- Which says "the question of what periodicity properties a not-necessarily-maximal n-bit LFSR may have does not seem to be addressed in the literature"

But from posts by others, it would seem that the particular implementation used by TWONKY had been widely-copied amongst other designs including being used in (somewhat ironically) a few 'Analogue' synths - Back in the 80's, I had considered building one of those that Maplin featured / published booklets on construction of one of these. But I only got as far as writing-out a list of all the parts needed, and never bought any PCB's to do this.
Therefore I'm interested in how much shorter than maximal length this is, and the effect on the sound of maxing it truly-maximal rather than wasting some of its 18 stages.


I found the wikipedia page on LFSR's to have a good summary on these - including a picture of a 31bit type, someone had built on breadboard which looks quite ambitious and a bit troublesome to fault-find if it didn't work: https://en.wikipedia.org/wiki/Linear...shift_register

And a few other useful articles, explaining these / providing tables of required taps / program-listings to implement these in software: https://www.eng.auburn.edu/~strouce/...6250/LFSRs.pdf
https://www.eetimes.com/tutorial-lin...-lfsrs-part-1/
https://datacipy.cz/lfsr_table.pdf
https://www.analog.com/media/en/tech...integrated.pdf
https://dsp.stackexchange.com/questi...shift-register
ortek_service is offline  
Old 26th May 2023, 11:21 am   #98
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Micro music box project.

I'm now in possession of an anatomically correct 4006 IC and have tried it and its companion 4070 both in the own-build PRBS generator and in the one Slothie provided for on Slothie-Ortonview. Unsurprisingly, using the right type of 4006 (CD4006) in circuit has improved the operation of both circuits and they now produce a random looking digital pulse train with the shortest 'bit' length being equal to the interval between two clock input pulses. As observed by Owen, I think, the circuit does often generate razor thin opposite-state spikes within the longer pulses which coincide with the falling edge of the clock / NADS pulse. Whether this will ruin our day remains to be seen.

Unfortunately, I'm away at the weekend as usual so I won't be able to pursue this further until next week at the earliest - Slothie can still overtake me on the home straight.
SiriusHardware is online now  
Old 26th May 2023, 3:42 pm   #99
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
Default Re: Micro music box project.

I think glitches on the output of the XOR gates has to be expected due to delays in the gates, for slow cmos at 5v they might be quite long. I’m not sure of the timing for sense or serial inputs on the INS8060, but fairly certain it will be long after NADS rising edge as the INS8060 first needs to decode the instruction. Maybe this is why NADS was chosen by the designer to clock the CD4006.
Mark1960 is offline  
Old 26th May 2023, 5:13 pm   #100
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Micro music box project.

The attached image shows a function generator generated active-low clock (upper trace) and a digitally stored capture of the resulting PRBS output (lower trace) which of course is constantly changing.

You can see a couple of the ultra narrow spurious spikes during the long 'low' on the far right of the image. They probably extend to full logic-one height in reality but are too fast for the sampling rate on my old Hameg Analogue-with-Digital scope at work to be able to capture in full detail.

As Slothie mentioned earlier the use of NADS rather than the system clock reduces the bit rate to one state per read in from the SENSE input and ensures that the PRBS output will not change state just as the uP is trying to read it.
Attached Thumbnails
Click image for larger version

Name:	PRBS_Output.jpg
Views:	40
Size:	58.0 KB
ID:	278794  

Last edited by SiriusHardware; 26th May 2023 at 5:33 pm.
SiriusHardware is online now  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 9:42 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.