5VCPC

[robcfg]

I'm interested in one for sure!


Sent from my iPhone using Tapatalk

[Bryce]

I'm interested in one for sure!


Sent from my iPhone using Tapatalk

My Order tracking template in Excel automatically includes your name at the top of the page

Bryce.

[1024MAK]

My Order tracking template in Excel automatically includes your name at the top of the page

Bryce.

Please add my name as well 

Mark

[pelrun]

Add one to the list for me, please 

[robcfg]

Haha Bryce, you know me too well!


Sent from my iPhone using Tapatalk

[Bryce]

I've re-visited the layout issues and I may have a solution to keep it mounted on the back of the drive. If I move the input capacitor to the back of the PCB it not only reduces the size of the device, but also acts as a standoff to avoid anything shorting against the drive metal.

Regarding the interference tests above, I should also mention, that this prototype doesn't even have a proper groundplane, so the end device should perform even better. I left the prototype running over night with a transient load switching between 150mA with 1mS 400mA spikes once per second to simulate bad case real usage. The device is still happily running this morning and still at room temperature I just need to check whether or how much 5V rail spikes it produces on a real CPC and then the design is more or less finished.

Bryce.

[Trixster]

I'm interested Bryce, please add me onto the list for 2 boards.


Andy

[Bryce]

Yesterday I optomised the layout further, mainly so that I can squeeze the device into the tiny space behind the drive on a 6128, but also to "balance" the PCB better so that it sits nicely on the back of the connector. To do this I had to move some components to the back of the PCB which also had the bonus effect of making the PCB even smaller. I've ordered some prototype PCBs to confirm the design and I'll be building them probably late next week. Here's what the (possibly) final design looks like. The board is now just 40mm x 24mm.

Bryce.

[robcfg]

It really looks beautiful!

[Bryce]

I've just ordered the parts for the mini-batch. The 5VCPC will cost €24. I will only be building 15 of these, so if you want one you'll have to be fast.

Bryce.

[robcfg]

That is without postage, right?

[||C|-|E||]

I think that i will ve taking one, if that is possible  .Yesterday I found my original CPC 6128 and, although the guy is pretty battered, I would love to fully restore it and install the adapter on it 

[Bryce]

That is without postage, right?

Yes, because postage will depend on where you live. However, it's pretty small and light, so postage shouldn't be much to anywhere.

@||C|-|E||: You'd already stated interest in one or maybe two on the first page. Is this an additional one or how many do you want in total?

Bryce.

[||C|-|E||]

Sorry! it is the same one i wanted to have. This memory of mine... 


If you ordered a small batch, one should be OK. This way more people will be able to have it 

[JonB]

I can use one of these. Put me down for it, postage to the UK.


Cheers
JonB

[Bryce]

Hi all,
     as the price of my 5VCPC has been called into question, I'd like to go through the design process step by step and explain why my device is not a $1 Chinese DC/DC converter and why my device costs what it does. For obvious reasons I don't have the scale of economics that the Chinese manufacturer will have, so I think everyone knows why it doesn't cost €1, but I'd like to explain the rest of the costs to you too. One point that needs to be explained up front is that a DC/DC converter is NOT the same as a digital expansion. It's not just the values of the components that are important, but the type, manufacturer and quality of the components all have a huge effect on whether the circuit works at all, disturbs all the electronics around it and runs stable and efficiently. On a typical digital expansion the designer will usually add a 100µf cap and maybe some 100nf decoupling capacitors. The type and manufacturer of these makes pretty much no difference, so you can choose the cheapest you can find and be pretty certain the circuit will work anyway. You can't do that with a DC/DC converter.

So here's the breakdown:

Mounting Location:
As you can see, the device is mounted on the back of the floppy drive, where there is exactly 10mm space. This obviously restricts many things and increases the chances of a short circuit. It's not mounted here because it looks fancy, or because I thought that would be cool. It's here because DC/DC converters transmit all sorts of crap that can disrupt and even crash the CPC, so it's important that the device is here hidden behind as much metal as possible and not right beside the mainboard. Unfortunately it also restricts what parts I can use because everything has to fit into this tiny space. The Chinese converter is considerably bigger and would not fit here.

PCB:
Due to the mounting position and shape a custom PCB was required. The PCBs cost me €5.80 each plus postage because I ordered just 15 boards which is considered mini-prototype volume. I could have ordered 100 PCBs instead of just 15 and the boards would have cost a lot less, but I won't sell 100 boards, so why order that many? I could also have got some chinese sweatshop to make the PCBs for a lot cheaper, but I believe in supporting local jobs, so the PCBs were made by a company whose manufacturing plant is in Ireland.

The regulator:
Why didn't I use a cheap adjustable regulator?
The Chinese devices use variable regulators which are relatively cheap (about €2 each). However, these have several problems:
- It needs to be trimmed / setup by the user to the exact voltage required and most users don't own a multimeter.
- They tend to drift with temperature so the voltage level is not always the same.
- I don't have space for a trimmer behind the drive even if I wanted.
- They are not optomised for 12V so their efficiency varies depending on what you set the voltage too.
- But most important: They have a failure mode, which involves shorting just one point on the PCB, which could easily happen in a space as tight as this. In this failure mode, the DC/DC converter can output more than 35V instantly, without warning. You don't have to be an electronics expert to know that this wouldn't end good for your floppy drive. Something that I definitely wouldn't sell or advise a CPC user to put in their computer.
Because of this, I went for a fixed voltage regulator, which (due to the fact that the IC contains precision resistors which an adjustable regulator doesn't), these cost quite a bit more. Not only that, I went for an older type that switches at a lower frequency than the more modern ones, again to reduce the noise being created. Using a newer type would have meant I could use smaller capacitors, but also mean that I would need to shield the entire PCB which I have neither space for, nor would it have been financially viable. In this case, the regulators cost me €5.50 each including postage as I could not find them locally and they are a relatively difficult part to source.

The inductor:
There are several grades of inductors available. You can use the small resistor type versions. These cost pennies, but they can't really handle the current required here. There are larger inductors wound on a ferrite core that can take the current and cost maybe 60 to 70c, but they are unshielded, so they transmit crap all over the place. I chose a shielded ferrite type, but not only that, I went up one step in the current capabilities which further reduces the amount of crap being transmitted and ensures that the inductor never gets hot. These cost €1.20 each, but they guarantee the least chance of any disturbance to your CPC.

The capacitors:
The capacitors on a DC/DC converter are pretty much the most critcal part of the circuit. To work properly, these should be high quality low ESR* parts. In fact many regulator datasheets will specify both the manufacturer and part No. that you should use. The Chinese devices slap on any old capacitors, not even low ESR most of the time and usually a lower value than is recommended. A typical standard capacitor will cost a few pennies. A low ESR capacitor will cost a bit more, then there are capacitors that offer a longer lifetime of up to 10000hrs and these cost more again. As any Amiga owner will tell you, capacitors don't last forever, so the longer lasting ones have a real advantage. I chose to go for a longlife, very low ESR part. Unfortunately these are usually 10mm wide barrels in the value I wanted, but I only had space for about 8mm as the PCB is 1.6mm thick and the total free space is 10mm. For this reason I also had to go for a slightly more expensive low-profile version.
The capacitor I chose in the end costs 95c each. A regulator datasheet gives a formula to calculate the value of the output capacitor. In this case the value came out as 560µf. However, I went for a larger 680µf capacitor because the disk drive 12V load is what's known as an inductive load and because it doesn't pull a steady current from the supply. This means that not only does the current get pulled in short bursts, but also that the current demand is out of phase with the voltage rise. This is something that DC/DC converters don't like and needs to be addressed. I partially counteracted this by increasing the value of the output capacitor. 
Although the input capacitor does not need to be low ESR, I decided to use one anyway. The reason for this is that DC/DC converters also cause spikes on the 5V (input) rail. These are going directly to the CPC mainboard. The 5V logic on the mainboard can live happily with up to 5.25V and will (for a while) survive with spikes up to 5.4V, however, if the 5V rail is going above this every time you access the drive, your CPC will most likely have died within a week. I have measured spikes of up to 5.45V on a chinese DC/DC converter. My device doesn't produce spikes above 5.15V. A further 100nf capacitor on the input filters out higher frequency spikes.

The schottky diode:
Another important component on the DC/DC converter. I could have saved some money here and used a cheaper SMD part, however the smaller SMD part would have produced more heat than I wanted, so I went for a large thru-hole part. Heat is important on this PCB as all three "heat producers": The regulator, diode and inductor are right beside the output capacitor and capacitor lifetime is inversely proportional to the temperature of the environment it is working in. The diodes cost 35c each plus postage.

User friendliness:
Most users probably don't own a soldering iron, nor do they want to be cutting wires or tracks in their CPC. For this reason I wanted the device to be completely plug and play with no soldering required. This meant adding the correct socket and plug to the PCB, which obviously a chinese converter doesn't offer. Although I was able to use a readily available plug, the socket is an obsolete part, that again I had to find an alternative source for and pay premium prices to get (€1 for socket and 32c for plug) and again pay postage for as it is yet again a different source.

The remaining parts cost just pennies and aren't worth mentioning. However if I may sum up:
PCB €5.80
Regulator: €5.50
Inductor: €1.20
Capacitors: €1.40
Diode: 35c
Connectors: €1.32
Other passive parts + solder etc: 50c

This is already over €16. It doesn't include some of the postage costs as I had to source the parts from 7 different vendors. It doesn't include the costs of the any of the other regulators / inductors / diodes etc that I bought to try out during the development. It doesn't include the costs of making the prototype PCBs seen in the earlier pictures. It doesn't include anything for my time and effort, nor for anything getting lost in the post or breaking.

As you can see, I put a lot of time and effort into ensuring that this device does what it should, does it safely and efficiently and won't destroy your CPC. My aim isn't to make cheap devices, my aim is to make high quality, good performing, safe devices for the CPC.

This may not be appreciated by everyone, but if anyone still feels scammed by the price I have put on this device I would be very happy to hear from them.

Bryce.

* ESR stands for Equivalent Series Resistance and more or less just means how easy AC voltage flows through it. The lower the better in a DC/DC converter which has lots of AC that it needs to be canceled out.

[robcfg]

That is a nice and detailed explanation, thank you!


Sent from my iPhone using Tapatalk

[Bryce]

So, I just got my hands on one of the Chinese DC/DC converters and thought I should share this data with you.
The following screenshot shows the spikes this device feeds into the 5V rail of your CPC if you were to install one in your CPC. In the first screenshot the DC/DC converter was supplying a constant 200mA load.

[attach=2]

As can be seen from the screenshot there are regular spikes of up to 5.6V on the 5V rail.

The next screenshot was taken while the converter was connected to a transient load switching between 100mA and 200mA to simulate reading a disk.

[attach=3]

The spikes have now increased to 5.68V

The TTL chips in the CPC are specified to work up to a maximum of 5.25V. I can't say for sure how long the CPC would survive this abuse. Not very long I suspect.

Bryce.

[CraigsBar]

I'll take one if you have any left in the 15 batch @Bryce

[yannis_uno]

Happy New Year & I am sorry for the late reply


I am also interested in one Bryce!


Thanks!

[Bryce]

Thanks guys, all orders noted.

Bryce.

[00WReX]

...if anyone still feels scammed by the price I have put on this device I would be very happy to hear from them.

Bryce.

Scammed...If anyone felt scammed then they can go design/source their own... nobody is forcing these on anyone. 

Cheers,
Shane

[Bryce]

Just for the record. If I thought for one second, that a $1 Chinese device was suitable and safe to use in a CPC, I'd be the first to open a thread explaining how to install it and wouldn't have gone to all the bother I've gone to.

Gesendet von meinem Motorola DynaTEC 8000X mit Tapatalk.

[chinnyhill10]

Let's cut to the chase here. Been reading this sorry tale and I'm going to say just this on the entire sad affair.


There are people who believe you can buy any old cheap shit and use it. Cheap PSU's, cheap car tyres, cheap meat, whatever.


The cheap PSU will be a fire risk that could damage your device. The cheap tyres won't handle as well or stop you as fast. The cheap meat could turn out to be horse or donkey slaughtered in an Eastern European country with poor animal welfare. Etc. etc.


The people who buy this crap always know better. "Oh I only paid 40 quid a corner for my tyres", "Oh I get my meat from that cheap supermarket", "Oh I found a PSU on Ebay".


They are idiots. Pure and simple. It's always about how cheap it is with no other thought.


To do things properly is never the cheapest way. Just because something looks OK on the surface doesn't mean it is. We're working with vintage hardware here guys. It's fragile. You don't need to be an electronics expert with a scope to know some cheap bulk buy from AliBaba isn't going to be very good for elderly circuitry which needs a nice smooth supply.


Bryce shouldn't have to defend himself or post pictures of his scope. You want to go stick some shit Chinese adaptor in the back of your CPC, go ahead. 6 months later you'll be whining on here for replacement chips for your broken CPC. Anyone who isn't a complete moron will buy the proper solution from Bryce because they'll know you need a smooth supply that doesn't put your hardware at risk.


Which is exactly the same reason why I wince at people using cheap PSU's on their CPC's compared to the more expensive units with proper data sheets from the likes of RS.


Components cost money, the software to design the circuits costs money, the equipment costs money. I know people who design and sell very popular lines for other micros. You have to be able to fund the components, the wear and tear on your equipment and have enough left over for your running costs and to go towards your next project. If you think anyone should churn out projects for just the cost of components then you are a fool.


That's my 2p to the forum 'drama'. I won't make further comment.

[Bryce]

One more thing that I think is important to mention. I'm not in any way saying that the chinese device does not do what it claims. It will convert 5V to 12V and is probably useful for certain situations, however, most DC/DC converters, even the more expensive ones are designed to give a good output voltage, the more expensive ones will give better results than the cheaper ones, but they rarely care about the input side of the DC/DC converter, for the simple reason that these converters are usually used in the following situation, where spikes on the input are pretty much unimportant.:

Source -> DC/DC Converter -> load (Sensitive electronics).

The situation in the CPC is completely different as there are sensitive electronics connected directly to the input side too:

Source -> 5V Sensitive electronics (mainboard) -> DC/DC Converter -> 12V floppy.

In cases like this, the spikes on the input become extremely critical as they can destroy the 5V electronics and need to be taken into consideration. I have looked for a generic DC/DC converter which claims to suppress voltage spikes on the input side in their specs, but I haven't found one yet. The main differences of my device involve improving the level on spikes on the input, the output voltage of my device is not all that much different than a generic device, other than that it was designed with an inductive load in mind.
The generic devices I've tested so far create up to 700mV of spikes on the 5V rail, easily enough to destroy a CPC mainboard. My device creates no more than 150mV on the 5V rail, even with an old crusty drive connected to the 12V side.

Bryce.