Some information can be recovered from the old forum.

[RGV250]

Hi,
you're probably right, I was thinking that as the Amicus18 was it would be the same, just as well I did not buy any.

Bob

[John Lawton]

...The flux I have is in a syringe but is awful for getting in the right place, what do you use.

I had a flux pen but the flux is very thin and runs everywhere. So inspired by the Youtube Macbook repairer Louis Rossmann, I bought a bottle of much more viscous stuff from Farnell, part no. 3215613

It's much better to use, I apply with a pointed object like a cocktail stick etc.

John

[top204]

The problem, for me, is that I do not have the capital to produce and, especially, advertise a development board.

Also, the Chinese flood the market with dirt cheap development boards and then the two buzz word companies also purchase cheap chinese boards and have their name on the PCB, with lots and lots of free advertising from magazines and internet sites, and over-used buzz words. So what hope does a small scale development board have?

Would it even cover the cost of the components on it?

[John Lawton]

Hi Les,

yes I can understand that. Neither do I, but I'm at a loose end at the moment so I thought I'd look at making a 16 bit dev board. First step is to assemble one of my Amicus24 boards and the necessary parts are arriving today.

Next stage would be to look at replicating the design or something similar. It would have to cover costs of course.

John

[top204]

There are a few things I would do differently on an Amicus board now.

Get rid of the FTDI USB/Serial chip altogether... They are stupidly expensive and were all that were available at the time of designing the Amicus boards.

Change the switch-over from USB to external voltage method. I have used a different method a few times now using a MOSFET and a Shottky diode. It does drop the voltage by approx 0.3 volts, but if used with the external voltage, this does not matter too much. I'm sure there are better methods out there for making sure the external voltage does not leak back up the USB and vice-versa.

Bring out 2.54mm jumpers to choose between 3v3 and 5v.

[John Lawton]

Hi Les,

agreed, except I've used the FT230X for a decade now in a product and they are reasonably priced. I can't find a stockist of the CH340E which I think is your alternative.

The Firewing board has a different power switching system, I might look at that sort of arrangement: http://www.firewing.info/downloads/schematics/firewing-r2.pdf

I know there are 'shields' available for the Amicus board format although I've never been into Arduinos so my other idea for a basic board design is something smaller that would plug into one of those solderless breadboards and not do much more than break out all the pins along the two main sides. Just the job for getting a micro up and running initially. Definitely have ICSP on board so having the right bootloader wouldn't be essential.

A bit like this, but for PIC24. https://www.mikroe.com/mini-pic32mx
Manual: https://download.mikroe.com/documents/starter-boards/mini/pic32/mini-32-manual-v101.pdf

John

[top204]

I've never liked those DIL type dev boards, because the SMD microcontroller is fixed and cannot be changed, and it needs a breadboard to fit it onto.

I still prefer a development board that things can be plugged into, and not-vice versa, and the microcontroller is a DIL type with a socket, and the USB has a socket and a socket for external power. So the board itself does all of the work standing on its own. i.e. An Amicus type board.

The switchover on the firewing, is the same switchover I originally used on the Amicus board, and was based upon the arduino board. But it is extremely inefficient, needing a comparator to switch on/off a MOSFET. At the time, I was in a hurry to get the amicus board out, but looking at it now, it was very lacking in its switchover, and some components on it, are simply not required.

I used the CH340G on the Positron boards because they are very inexpensive and use a standard 12MHz crystal, and the drivers are suitable for windows 10 and 11. If board space is not a problem, then a very inexpensive HC-49 crystal type can be used.

[John Lawton]

Agreed, a DIL socket is very desirable.

The adjustable regulator and schottky diode on the Amicus24 is a bit clunky but I like the comparator idea if it gives a definitive switchover. A comparator can actually be cheaper than a power FET.

I still have trouble finding stock of any CH340G devices.... where to get them?

John

[Gamboa]

John,

Here you can buy them. It is a reliable site.
https://www.lcsc.com/product-detail/USB-ICs_WCH-Jiangsu-Qin-Heng-CH340G_C14267.html

Regards,
Gamboa

[John Lawton]

Hi Gamboa,

many thanks for that, I hadn't come across that company before.

Now to find an English datasheet....

John

[EDIT] found a datasheet (actually several).

[top204]

If memory serves, I added the Shottky rectifiers to the Amicus24 circuit, so that it could also be powered by an external 5 volts going to its 3v3 regulator, and no current would creep back up the 5v regulator. I had to use the variable voltage regulator in order to cater for the voltage drop of the Shottky rectifier, so the circuit still saw 5 volts, but that was a long, long time ago.

Below is a simple circuit I have used in various formats in some projects to switch over seperate voltages. It uses two P-Channel MOSFETs, so the driver MOSFET can drive the power MOSFET so it can carry more voltage that the drive can give to its gate. As you can see by the circuits, when external power is applied, it cuts off the voltage from the USB, so the driver MOSFET does the same as the comparator chip, but is a lot smaller and easier to do. The circuit will also switch off the USB power with lower external voltages, depending on the driver MOSFET type.

Again, if memory serves, the current when the circuit is in the off position is also very small, but this relies on the resistor values and the MOSFETs used. Some MOSFETs will operate on higher value resistors. Forget the values of the MOSFETs in the circuit, they were used as general purpose P-Channel types for the simulator:

   

[John Lawton]

Hi Les,

yes, power supply arrangements can be quite tricky as you noted. Thank you for the simulations, I'll replicate them in my TINA package - which incidentally, seems to do digital simulation including PICs (but I haven't tried it out).

In any case I don't think a lot of the overall component budget would go in this area. The CH340 device however saves quite a lot over the FTDI one. I see that some models of that range don't need crystals which is nice.

As I understand it, vitally the CH340 must have a DTR line to enable a PIC reset in order to trigger a bootloader.

John

[RGV250]

Hi John,

As I understand it, vitally the CH340 must have a DTR line to enable a PIC reset in order to trigger a bootloader.

I think ICSP is a must?

Regards the footprint, I think the Amicus/Arduino footprint is good as it does allow for quite a few shields which can be just dropped on for testing.

Just my thoughts.

Bob

[trastikata]

Hello Les,

Since most of the PICs now run at 3.3v, if I may suggest instead of designing with 5v USB and 10v supply, think of 5V USB and 4.2LiPo battery input, both going to a 3.3VLDO regulator with a 100mV dropout or so.

Another thing to consider is that the 5V USB is almost never 5V - what I do is using a diode to drop 0.7v and then ME2108A50 as a simple fixed voltage step-up IC, really cheap but good IC.

[RGV250]

Hi,

if I may suggest instead of designing with 5v USB and 10v supply, think of 5V USB and 4.2LiPo battery input

I am assuming this would be for a development board rather than a finished project so probably just the USB supply as you will probably have it hooked up to the PC. If not then I would think most have phone charger packs which have a USB output. Not having the other jack would free up space.

Another thought is for Grove / Quiic connectors as they have some nice modules.

I wonder if John is serious about this perhaps a new thread and see how many people would be interested and what they price they think they would pay.

Bob

[John Lawton]

Hi Bob,

interesting about GROVE and QWIIC etc, I hadn't heard of them.

This is interesting on I2C interfacing: https://hackaday.com/2022/05/04/the-connector-zoo-i2c-ecosystems/

Agreed, a new thread is called for now, I'll start one shortly.

John

[top204]

The external voltage of 10 volts I used in the simulation was just for a simulation. The driver MOSFET will switch on the switching MOSFET at much lower voltages. Unlike the comparator, that has to detect a voltage over a specific value based upon its resistors. Even when using 5 volts from a USB, that can vary slightly, it does not effect the microcontroller too much because any ADC working within it can use the FVR (Fixed Voltage Referenece) peripheral, and this is always a better method than using the VDD as +Vref, because it gives a quantization value that fits better with the bits of the ADC. i.e. 4.096v, 2.048v or 1.024v.

What the switching MOSFET does is stop the USB voltage coming into the circuit and feeding the regulators, so the external voltage will not creep back up the USB and hurt the USB transceiver in the PC. It also does not drop the voltage too much when switched on, as long as the MOSFET has a good On resistance, and is switched fully on. But this is down the the perticular MOSFETs used and their gate resistors. MOSFETs are very peculiar devices indeed, and I am still learning about them, even after all of this time. And still get things wrong when using them. :-)

I think this is where david would be able to help more, because he seems to have a much deeper understaning of the innards of electronics.

I found that the CH340 USB to serial chips with internal oscillators are quite expensive, and are not much different in price to FTDI devices, but a CH340G is dirt cheap, and even with an external crystal and two small capacitors is far less expensive.

[John Lawton]

P-channel FETS, easy. Think like valves, but using positrons instead of electrons and turned upside down.



John

[top204]

If only it was that easy John. :-)

I've actually killed high amp MOSFETs because they were switched on at the wrong time in a cycle. I think it is something to do with high power current drain and gate going high/low too fast or too slow, or something like that. On one project, I actually had to ease the current drain until I switched on the gate, then it operated perfectly and never failed.

Also, unless it is a logic gated MOSFET, a higher voltage switching cannot be done with a lower voltage on its gate. This is something I learned the hard way, many years ago, after I had killed few MOSFETS because they were not quite on and causing a higher resistance between the Drain and Source, so they get very hot before they pop! So I always use a smaller MOSFET driver type method of switching now, unless the same voltage is being switched between the Draind and Source that will also be produced on its gate.

But MOSFETs with internal drivers are still more expensive than standard MOSFETs, even when it takes two of them.

[kcsl]

Just a suggestion, but is a development board really the way to go these days?

It depends on who your target audience is. If it's education then I would agree that dev boards are probably a good approach as lessons can be tailored around what the dev board can support, but for hobbyists or engineers I'm not so sure, unless you are looking to compete with Arduino, which will be very difficult now I think.

The problem with them is they are typically tied to one type of PIC and have a fixed number of hardware resources. They never have enough LEDs or the right size keypad, or don't have the 40x2 LCD you want to use or don't have a specific gizmo that you want to experiment with. You can also end up spending money on hardware you will never use. Do I really need an electric motor with a fan on it.
Yes, you can connect the dev board to the gizmo but now your moving into breadboard territory.

Typically the only advantage they give you is a bootloader that means you don't need a dedicated programmer. However, if you are ever to do anything serious with PICs you will probably need a programmer at some point.
The pictures show one of my carriers for a 40 pin PIC. Drop it on a breadboard, give it power and plug the programmer into it and away you go. I've a stack of these carrier boards for everything from 7-segment displays, to complete serial interfaces (MAX232, 9pin D-type, patch jumpers etc). Anything that I use repeatedly I create a carrier for. I've got a carrier that has 4 74595 shift registers on it. Anything to mean I can prototype quickly and not have to keep looking up which pins do what.
The concept could easily be expanded on to make it more robust for beginners if required.
My PIC carriers for example have provision for a diode in-line with the power pins; this means that the programmer doesn't attempt to power the entire project via the ICSP. A USB interface could easily be added so it has bootloader support and power.

Doing it like this means that all the end user needs is a breadboard, a PIC module with bootloader and they are off. They can add things easily and don't buy hardware they don't need.

The 3rd picture is for a project I'm working on. My PIC carrier with programming socket just drops into the project. Once it's all debugged and working I will remove the carrier and drop the PIC straight into the project so I can reuse the carrier.

I of course can only speak from my personal experience which is limited with dev boards. The only one I ever had was the TK3 I built from Everyday Electronics, and once I'd flashed the LEDs it only ever got used for it's PIC programming abilities.

Just my 2p worth.

Joe