Building Frank C's Hedghog 625 PAL to 405 System A standards converter.

It's time to big up someone else's project.

Enter Frank C's Hedghog converter...

Details of the project an be found here.. http://electronics.frankcuffe.ovh/hedghog

It's a stunning project, and features a few useful functions not found on my other standards converter, the Aurora SCRF405 (which for those of you who don't fancy the DIY approach, can be found here http://www.tech-retro.com/Aurora_Design/Single_Converter.html)

It consists of a digital video converter (The TVP5150) which converts PAL to an 8-bit ITU−R BT.656 format. These one's and zero's are fed to an EP2C5 FPGA development board, which mounts to the top of the PCB. The magic happens in here, and the resultant 405 line system A video output emerges from the resistor ladder DAC. This is then passed to a MC44BS373 modulator IC, which is tunable on all Band I and Band III channels. The tuning of the modulator is also taken care of within the FPGA, and is set with a hex switch. There's a second MC44BS373 which is modulating audio at the required 3.5MHZ spacing below the vision carrier.

The unit produces aspect ratios of both 4:3 and 5:4 (5:4 being in use up to 1950)
There's a Pedestal function which lifts black level 50 mV above blanking level. This was abandoned on transmitters post-war, as it reduced the efficiency of the transmitters. It may help with early sets that suffer from flyback lines.
You can select between normal broad pulses or broad pulses with equalising pulses. Equalising pulses were never in the spec, although it may help interlace on some sets (ironically it upsets interlace on some, and causes the top of the picture on some Bush sets to "hook")
The three-line interpolator has three settings: soft, medium and sharp interpolation apertures.
Switchable 1KHz or 400Hz test tones.
There's also a PM5544'esque test pattern, and stair-steps.

These functions are all available via switches on the front panel.

PCBs were ordered from the most-excellent pcbway.com, and arrived within 7 days.

The FPGA board came from eBay.
The MC44BS373's I had to source from AliExpress, as it's now classed as obsolete, however, at the time of writing this, there appears to be plenty of stock from "grey" suppliers. I dislike doing this, as many times I have been caught out with fake parts! Caveat Emptor. I used the AliExpress supplier YT Electronics components co.,ltd


The phono sockets also came from AliExpress (link) and fit with a minor modification.

Assembly is straight forward, don't let the surface mount put you off. Get a decent quality gel flux. If you've seen my videos, you may have realised I suffer with a benign tremor (nothing to worry about, I've had it since I was 14), and I can mount this stuff with ease. I do have the advantage of a microscope.

A bit of blue-tac helps hold the board in place. First off I fit the semi's.

Then the passives...

Programming the FPGA is straight-forward in windows, using the Quartus II (13.0 sp1) software, and the "USB-blaster" supplied with the FPGA board. I failed to get the software to work under Ubuntu (although it would "see" the programmer hardware, the program option remained unavailable). When correctly programmed, there's a binary counter running on three LEDs on the board.

 Sockets and switches are mounted.

And the FPGA board fitted.

... and it's switched on!











However, some fault finding was required, as although the test patterns and test tones were OK, there was no converted video. With the help of the designer, Frank and the VRAT forum (here), it was apparent something was wrong with my FPGA board. After some fault-finding, a tiny solder bridge was found on the FPGA board, shorting out two of the lines. Once this was removed, the converter worked faultlessly.

Moire pattern is being caused by the camera, the actual picture quality is superb.

I like the additional functions available from the front panel. Picture quality appears to be on a par with the aurora converter.

What I'd really like is a centre-cut out function for dealing with a 16:9 input.... Frank?

The best laid plans of mice and man... Fellows A75 laminator continued ...

If you've been following this for a while, you'll know I modified a Fellows A75 laminator to get a more reliable way to make toner transfer PCB's. (You can find the original articles here and here.)

As I had predicted, we were now using the item way outside of it's specification, and that's it's life expectancy was certainly going to be, erm, compromised....

Well, tonight it happened. It stopped getting warm. Now I've got a commercial prototype to get out, and I need it. Let's see what's happened.....

So, the poor thing is once again disassembled. Those two red wires you can see running from the other side of the PCB are connected to the element.... A quick check with the meter, and ... it's open circuit. Damn.

Now, in the last modification I decided to raise the temperature by modifying the controller. So what if there's one of those non-resetable thermal fuse things on the element somewhere....

And there is! Mounted against the lower heating element...

It's duly removed, the two wires soldered together, and insulated with a piece of heatshrink...









Back in business!

Now, we've removed an essential safety device, so I'll re-iterate my previous warning... We're putting paper in here, which is being exposed to higher temperatures for more time than it would be in normal use.... If it should jam, there's a risk it could smoulder and catch fire, and ruin your day... DON'T USE IT UNATTENDED!

... you nkow what , I may just build a combined speed/ temperature controller for it, and have done ...

More modifications to the Fellows A75 Laminator for toner transfer.

If you haven't seen the first bit of this ... click here.

Made quite a few boards with the laminator and toner transfer method recently, and results have generally ranged from acceptable to good.

There's the occasional fail, but I want to improve repeatability.

I've made some measurements of the temperature of the rollers, and it's temperature controlled at about 100 deg C, and sometimes, the board just doesn't get hot enough to melt the toner and cause it to transfer. We need to warm things up a bit....

I removed the PCB and reverse engineered enough of it to work out how the temperature control is done... IT'S IMPORTANT TO REMEMBER THIS CIRCUIT IS "LIVE" AT ALL TIMES, AS THERE'S NO TRANSFORMER. I used a mains isolation transformer whilst making measurements and working on this unit.

Mains comes in at JP1. Mains is dropped via a capacitive dropper C1 (R2 prevents the cap remaining charged up once the unit is unplugged, and removes a shock risk) R1 provides a bit of current limiting, D1 and D2 rectify the output, which is stabilised by ZD1, and C2 is used to smooth the supply. N3 is a 5 volt regulator, and C3 provides a reservoir for the 5v rail. The temp sensor looks like an ordinary 1N4148 diode.... it may well be. It's pressed up against the lower rubber roller. When it's at room temp, it's got about 4.2V across it. When it get up to temperature, there's about 2.4V across it. This is loaded by R32, and connected to the inverting input of N2 via R7, and back to ground (neural) via R8. A voltage reference is provided by the chain R9,R11,R12 and R13 (approx 2.6V) and applied to the non-inverting input, forming a comparator. Stability is provided by some feedback provided by R6. The output of the comparator is fed to the gate of a triac, T1... Now a note on D7... it connects to N1, which the manufacturer has very kindly filed the number off.... I imagine it's some sort of micro controller. It most likely provides some timing (the unit shuts the motor off after 30 mins) and some zero-crossing pulses to D7 so the circuit can phase modulate the triac. I've not illustrated this part of the circuit, as it's not relevant, and difficult as I don't know exactly what's going on inside the mystery N1. 

OK, so we need to alter the voltage reference at pin 5 of N2. I remove R9, and place a 10K pot in it's place.

I adjust the pot until the temperature stabilises at 130 deg C.

The pot's then disconnected and measured, and it's 3.6K, which is rather convenient, as that's a preferred value! R9 is replaced with a fixed 3.6K part.


Tests prove transfer is definitely more "robust". The toner seems thicker, and more difficult to remove.

We're now operating this unit way outside of it's design spec (as if it wasn't bad enough before!), so don't leave it operating unattended!!

Arduino programming tool. ISP and IDE.

The big hifi pre-amp is on hold for a bit, as I await some parts to arrive.

In the interim, I've wanted to make a programming tool for ATMEGA328 chips, that will ease programming changes when using them in stand alone units. I've looked about on the net, and there's some solutions which are nearly what I want... but nothing quite hits the mark.

The idea is to have a ZIF (zero insertion force) socket into which the blank chip to be programmed is inserted. I want it to be able to program chips with the bootloader, as well as straight from the IDE.

The fitted micro is loaded with the "Arduino as ISP" sketch. The switch, S1 & S2 (actually 1 four-gang switch, I just didn't have a library for it!) is used to switch between ISP programming for bootloader, and IDE Programmer. LEDs indicate various functions. D1 is ISP heartbeat, to show it's OK, D2 is Error in ISP programming, D3 is ISP programming activity, D4 indicates ISP mode, D5 indicates IDE mode (also lights when in ISP mode, as I forgot about the steering diode D5), and D6 indicates slave activity during bootloading (ICP1).

...a simple board is designed and etched....

The keen-eyed amongst you will notice the steering diode, D5 is missing from the layout. I mounted it underneath the board, as I'd missed it...

... and it works a treat !

Oh yeah, a quick note on R2 and R7... both 1 megohm across the crystal. I've picked up this bad habit from the original Arduino design. It doesn't need to be there. It was originally to provide some bias for gate-based oscillators. The ATMEGA328 has bias provided internally. It's a waste of 1 meg resistors! Don't bother to fit it, and I promise never to include it again!!

Tip for drilling homemade PCBs.

Here's a quick tip .... if you're as myopic as I am, try this out...

Here's my PCB drill in it's stand. It's a foul thing. It was dirt cheap when I bought it 15 years ago ... The rubber coating on it has gone all sticky ... yuck...

Now look at the bottom...

I've mounted one of those small 3w LED's there, left over from another project...










Makes alignment of the drill bit with the pad on your PCB a breeze.....


You'll notice it's slightly off-centre, this is avoid drilling through it ....

.... yeah, you know how I found that out!

Modifying the Fellows A75 A4 laminator for making PCBs. Toner transfer the posh way!

Now, if I'm making something up, it's often easy to make a simple single-sided PCB up quickly, and professionally by using the toner-transfer method of PCB manufacturing. I've been doing it for years, but the results are always a bit hit-and-miss. Use google if you haven't heard of it before, there are many guides on the web (although no two the same!) , and I don't intend to duplicate those here.

I have tried many different types of transfer paper over the years, from the proper stuff, to plain old A4 paper, the best I found for years was a copy of the New Scientist magazine! Sadly they changed to thicker, shiney paper a few years ago, which doesn't work well at all. The best at the time of writing seems to be "The Economist", which is a shame, as it's nowhere near as good a read as "New Scientist" before it get's recycled into PCB making materials!!

... anyway ...

The last board I did, I tried an experiment, by borrowing Mrs Doz' laminator from her office. I idea was to see if it would transfer the toner from the paper to the board, whilst keeping a constant and steady pressure on it. Well, results were mixed... it really pushed the board through too fast for proper transfer to take place. I ended up transferring the toner with the conventional clothes iron, and then passing the board through the laminator, which resulted in an improved board after a couple of goes. It did, however give me an idea ... make the motor run slower....

So, rather than incur the wife's wrath and modify her laminator, I picked up a Fellows A75 A4 laminator from Argos for £16....

I initially tested it would take the thickness of the board by passing a blank piece of PCB stock through it... yep, works a treat.

So, to disassembly... There are 4 pozi screws underneath (great, not even some torx or bizarre security screws to bother me!)

So (from left to right) we have an essential mug of tea, the motor, drive rollers and heaters, and finally a PCB on the right....

A couple of things to notice about the PCB... 

There's some sort of timer in there to shut the unit down after it's been on for 30 mins (just in case you leave it running). It's unlikely to bother us much....

The low voltage electronics (the timer, and LED's etc) are driven from a power supply formed by a wattless capacitor dropper... so a word of warning... all of these voltages on this board will appear to be referenced to the mains. This would present us with a shock hazard and a problem if we want to interface to them...

On to the motor....

It's something I used to refer to as a "Le Cruzet" motor. A synchronous motor with built in gear box. As we can see from it's label, it rotates at 4 RPM (4.8 RPM if you live on that funny bit of the planet that uses 60Hz mains). That's a bit of a pain. I was hoping for a small DC motor I could drive with some PWM to get the speed nice and regulated (although if I was building a cheap laminator, I'd be using exactly this motor too!) 

So, how can we control it's speed?

Well, it's synchronous, so we'll need to lower the AC frequency driving it .... Ugh ... don't fancy this idea much.. generating a few tens of hertz at 240 V RMS is not going to be a cheap fix....

What about phase modulation? Daft idea. It may vary the speed, but will destroy the motor's torque, and won't vary the frequency. May cause overheating and failure in the motor's windings because of the resultant unpleasant waveform.

What about PWM? Nope, DC only ..... or is it????

How about really, really slow PWM? Really really slow.... 

I'm talking about switching the AC to the motor on for ~.5 second, off for a second, and so on ......

It'll inch the board forward, cook it for a bit, and inch it forward again ....

It's just stupid enough to work....

So what do I need? 

A 555 timer, a slack handful of passives, a transistor and a relay...

or an arduino (or similar, running a modifed blink sketch!),a transistor and a relay...?

I'll go with the latter, as it'll be easy to adjust by fudging the software, and most of the hardware is built for us.

Enter the relay module, and the weird and wonderful Digistump Tiny85 module... just the ticket, and small enough I should be able to squeeze them into the case.

I've used the normally closed relay contacts, so if I do want to use the laminator without the speed control, it will work fine with the speed control switched off. The relay contacts are connected in series with the motor, and the relay control pin to P0 of the digistump.

A design is laser printed onto a piece of paper from "The Economist", and a test piece of board stock is cleaned up (I know they're a different size, but it's only a test)

The paper is tightly folded round the board so it doesn't move, and stuffed into the mouth of the hungry laminator! The "printed" side needs to be facing the bottom roller, as that appears to be the roller that's heated.

I left the speed controller off (so the unit was running at normal speed) whilst it reached temperature.

I ran a couple of tests, trying different speeds.




and into the sink!













I finally settled on 4 seconds off time, and 0.5s on time.... which produced nice, repeatable results ....

and left us with the digistump code looking like this....

void setup() {
  pinMode(0, OUTPUT);
}

void loop() {
  digitalWrite(0, HIGH); //Turn the motor off
  delay(4000);
  digitalWrite(0, LOW); //Turn the motor on
  delay(500);
}

So the electronics is all boxed up, and driven from a redundant mobile phone charger...

A few words of warning....

We're putting paper in here, which is being exposed to high temperatures for more time than it would be in normal use.... If it should jam, there's a risk it could smoulder and catch fire, and ruin your day... DON'T USE IT UNATTENDED!

It clearly states "Up to 80 Micron" on the box. We're really taking the P*$$ here, as my boards are 1.6mm thick, plus the paper (folded) so don't expect it to last ... but while it does .... get making!









... oh, and you can use the same idea to print a "slik screen"...

Metcal MX-PS5000 repair.

Been very busy, in a commercial sense, building some transmitting equipment ...

... and just as I'm about to solder the last two wires on the front panel of said equipment, the iron goes cold ...

The soldering iron in question is the very, very wonderful Metcal MX-PS5000 system. I used to think Weller made the rolls-royce of soldering irons, but these things are just in a different league!

They heat the element by RF induction heating. 13.560 MHz to be precise. It's got all manner of error detecting, it'll put itself to sleep when not in use, and also has outputs for two irons. It's also expensive.

So I think, it's probably gone to sleep ... nope.
Perhaps the tip has failed? No... (takes two seconds to change too!!)

Damn ... the emergency iron is flagged into service, and the job finished...

A new MX-PS5200 is ordered (at about £500!.. ouch). We simply can't be without it.

But let's see if we can't do anything about the old one...


Firstly, remove the retaining nuts from each of the F-connectors.


















Flip the unit over, and remove the 4 security torx screws, one in each corner.

Right, that's got it out. To remove the board there are several small screws to undo, and remove the screws from the heat sinked (heat sunk?) devices, noting which are insulated from the heatsink, and which are not. The wires connecting the power supply in the back and the main board are quite short.















First thing was to check the presence of the +48V  and  +24V supplies from the switched mode mounted in the rear third of the enclosure.. all present and correct... damn.. out goes any ideas of a blown fuse and a quick-fix! There's no hope of getting a schematic, so it's down to the poke-and-hope method of fault-finding....

Checking a few parts, shows the main output FET to be dead short ... It is (or rather was) an IXFH12N50P, which is no longer available. An IXFH16N50P looks like a good candidate, and is pressed into service.... still not warming the iron... damn. It's gate is driven by a drive amplifier, which is in turn driven by an oscillator formed by a crystal, and a 74HC04 hex inverter... the hex inverter is duff, and is replaced, and so is the supply to the drive amplifier...

It's an LM2576-ADJ (not that you'd know from the photo) .... and replacing that restores the supply to the second stage of the drive amp, but not all of it .... damn.














The final piece of the jigsaw is D12 in the small switched mode buck converter .... it too was dead short ... lord knows what the original is, there's no circuit anywhere on the web, so I selected a 1N5819HW for the job, based on the datasheet for U3. It's physically smaller than the diode it replaced, but it seems to work! I've got a hot iron!













Excellent ! We now have a spare in the workshop.

Not sure I like Metcal's new logo much ...

Arduino Audio Wattmeter.

Many, many moons ago I needed something to measure audio power in a hurry. So I lashed up a quick and dirty audio power meter.

It was nasty....

... very nasty, but did the job at the time.

It was simply a switchable 4 or 8 ohm load (actually 4.7 or 8 ohms)  which was rectified, and used to drive two meters, each measuring full scale at 10 watts. The fan was driven from rectified audio. Most unpleasant.




Anyway, the load was good, as was the fan, but I needed some sort of better instrument, so a plan was dreamt up. Measure the peak voltage across the load, do a bit of maths and get the results displayed on a 1602 display, driven by an arduino. Great ....

So audio comes into the load (R1 & R2), and is rectified by a bridge. There's a small bit of smoothing on the output, so we can get a steady-ish DC. This is divided by a voltage divider, R3 and R4 and the resultant is fed to the arduino's A-D converter. Couple of things to note here, the load is switched, either 8.0 Ohms, or 4.7 Ohms, by S1. This also lets the arduino know which load is selected. S2 is a "BTL safety switch". It completely isolates the left and right channel. It also stops the arduino reading the right channel. If you're testing a BTL(bridge tied load) amp, or any type of amp which may not take kindly to commoning up of the negative poles of the speaker, best click this across. It also grounds the second analogue input to prevent spurious readings. A PWM signal is output from D6, and is used to drive a FET to run a fan. I chose a stupidly large FET here, you can pick something more sensible. I just happen to have a bucket of them. 

This will only give accurate results when feeding in a sine wave.

That's it really....

 Dummy load salvaged from the old unit...

Front panel coming together.

... sizing it all up....

... starting the wiring...

Looking good until I noticed that I'd mistakenly picked up a 16x1 display instead of a 16x2!... and the 16x2 requires a slightly smaller opening. Drat! ... some hot melt later will have to do (and it was all going so well ...)

Rectifier board and PWM FET

Rectifier was a little overkill! Still I had two identical, salvaged from a switched mode PSU.











Setting up ready for calibration...

Apply 28.28VDC accross the outputs of the two bridge rectifiers, and with the unit set to 8 Ohms, adjust each pot until the display reads 50.0W. Easy... (Why 28.28V ? 20 VRMS into 8ohms is 50Watts. 20*1.414 = 28.28Vpk... see here for more)





All calibrated and ready for action!
Top line on the display reads left channel, bottom reads right. The three readings are instantaneous power, average and peak.











And last, but by no means least... the code....

/* The "I'm sick of that analogue meter" Audio wattmeter (c) A.G.Doswell 2nd April 2016 License: The MIT License (See full license at the bottom of this file) V1.0 Arduino measures the DC output from a diode bridge across the load resistor, does the maths and displays the results. Peak-hold feature, averaging and PWM fan control too! This is only going to give true readings with a sine wave input. VRMS = VPk + 1.41 / 1.414 Power = VRMS^2 / R R is either 4 or 8 ohms (switchable) */ boolean rPin; int sensorValue; float voltagePk; float voltageRMS; float powerR; float powerL; float loadR; float peakR = 0.0; float peakL = 0.0; int loopCounter = 0; float averageL = 0; float averageR = 0; int averageCounter = 0; float averagePowerL = 0; float averagePowerR = 0; int fanPWM = 0; int fanPWM2 = 0; #include <LiquidCrystal.h> LiquidCrystal lcd(12, 11, 5, 4, 3, 2); // initialise the LCD void setup() { pinMode (7, INPUT_PULLUP); // switch connected to 7 defines load resistance, Open for ~8 ohms, short to GND for ~4 Ohms. lcd.begin (16, 2); } void loop() { rPin = digitalRead (7); if (rPin == false) { loadR = 4.7; // My 4 ohm load is actually 4.7 ohms } else { loadR = 8.0; // .. but my 8 ohm load is spot on! } sensorValue = analogRead(A0); // Output from Right channel connected to A0 voltagePk = sensorValue * (35.0 / 1023.0); //Convert the analogue reading to a voltage voltageRMS = (voltagePk) / 1.414; // Vpk to VRMS powerR = (sq(voltageRMS)) / loadR; // VRMS and load to power if (powerR > peakR) { // update peak value peakR = powerR; } averageR = averageR + powerR; //running average sensorValue = analogRead(A1); // Output from Left channel connected to A0 voltagePk = sensorValue * (35.0 / 1023.0); //Convert the analogue reading to a voltage voltageRMS = (voltagePk) / 1.414; // Vpk to VRMS powerL = (sq(voltageRMS)) / loadR; // VRMS and load to power if (powerL > peakL) { // update peak value peakL = powerL; } averageL = averageL + powerL; //running average fanPWM2 = (powerR + powerL) * 8; // get a value for the fan speed PWM output if (fanPWM2 > fanPWM) { // if it's greater than the current PWM value, then up date it fanPWM = fanPWM2; } if (fanPWM > 255) { // limit the maximum PWM value to 255 (flat out) fanPWM=255; } analogWrite (6, fanPWM); // write it to the if (averageCounter == 100) { // sets the Average reset averagePowerL = averageL / averageCounter; // create the two averages from the running value averagePowerR = averageR / averageCounter; averageCounter = 0; // reset the counter averageL = 0; // reset the running averages averageR = 0; fanPWM --; // decrement the PWM value for the fan if this are cooling down! } if (loopCounter == 1000) { // when this has run 1000 times reset the peak value. loopCounter = 0; peakL = 0; peakR = 0; } lcd.setCursor (0, 0); lcd.print (powerL, 1); lcd.print ("W "); lcd.print (averagePowerL, 1); lcd.print ("W "); lcd.print (peakL, 1); lcd.print ("W "); lcd.setCursor (0, 1); lcd.print (powerR, 1); lcd.print ("W "); lcd.print (averagePowerR, 1); lcd.print ("W "); lcd.print (peakR, 1); lcd.print ("W "); lcd.print (loadR); loopCounter++; // decrement the counters averageCounter ++; } /* Copyright (c) 2016 Andrew Doswell Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permiSerialion notice shall be included in all copies or substantial portions of the Software. Any commercial use is prohibited without prior arrangement. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESerial FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHOR(S) OR COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */