usbasp firmware that generates a 1mhz clock signal

[rperkins]

I was thinking about people whose fuses get mangled on their boards.
I know the arduino can generate an external clock to revive them.

I think I have seen usbasp firmware that generates an external , relatively slow clock, but cant find a link.
Have you guys seen a link for something like that. ??
if not,
would it work.?

I'm thinking to generate a 500khz or 1mhz clock signal on an unused pin. USBASP has an unfinished RX/TX line on PD0/PD1. The programmers I sell just terminate those lines to through a resistor to ground but you could solder a wire on them and run it to the 9x/9xr board. Would that be a strong enough clock signal to drive the clock on the 9x/9xr. Pretty sure at least one of the chinese non usbasp ( usbisp s51 , progisp, etc) software interfaces to their hardware have an option to generate an extra clock signal.

thanks

[jhsa]

I did test a 1Mhz clock on my arduino.. it looks good on the scope, so I guess it works..

[rperkins]

Thanks for the reply.
It was Nick Gammon's sketch 'atmega hex uploader' that generates a clock (around line 1050)
https://github.com/nickgammon/arduino_s ... loader.ino

I think this would be a good feature for usbasp hardware. Not sure on the best way to implement it . Which pin ? Always on ? What speed ? Etc,etc

[Flaps 30]

Any unused pin on the ten pin connector that doesn't clash with what is going on with the six pin variety for the 9xr

Always on as it will only be used when connected to the clock pin of the target.

Speed.. Hmm.. Don't know that one without looking at the data sheets and the like. No doubt someone has the answer to that.

[MikeB]

Pin 3 of the 10-pin connector is the only one documented as "no connect". The rest are either signals, power or 0V.

Mike.

[rperkins]

pins 4 and 8 of the 10-pin connector were designated for TX/RX by the fischl.de schematic. You're right , most everyone grounds those pins now. Those pins go to PD1 and PD0 on the MCU. It'd probably be wise to see which pins can be run from a hardware output compare pin before we narrow it down too much. Units that provide the "slow SCK' jumper, that may be a good pin as the -B option makes that jumper mostly obsolete.

Ideally I wanted to look at the most popular usbasp implementations and generate a change that would work with all of them. However, everything is in a state of flux right now. The most common usbasp are HK and betemcu.cn ? HK retired the old working model, and has suspended sales of the new model. Just found out today betemcu.cn is switching over to a single sided board also. Was told today 'USBASP sells so good this month and the factory is product the new batch' . Now to me that shows the effect of HK usbasp sales going offline. So who knows what will be most common implementation in the near future. I do believe betemcu.cn is out of the units we have come to know and shipping the 'new' model.

I dont think it is asking too much for the end user needing a clock signal to need to solder a wire to the board. The existing betemcu.cn had typical( I can never remember the numbers) sized smd resistors ( to ground) on the above mentioned TX/RX lines. I was thinking remove the resistor and solder on the pads. But now, I fear they are moving towards smaller SMD parts, as best I can tell, which may be too small to solder.

If that unused pin 3 went back to the MCU that would be great.

Thanks for the information and input.

[Kilrah]

I really wonder why HK stopped selling the "old" model. It worked perfectly well for "normal" applications, and had both 10-pin and 6-pin plugs... pretty much the best around.
Only the special cases needed changing FW just like most others.

[rperkins]

IDK. I would expect the cost of a single sided board would be cheaper.

HK did reduce the price but not by much more than the price of the 10 pin IDC and the labor involved to cut and lace the 10 conductor ribbon cable into the 6 pin IDC. I looked into making my own 'combo' cables and determined it was just cheaper to buy the adapters. To be fair, what I figure for labor is higher than their cost.

Touched upon the lower cost of a single sided board today. I think the reply was 'it's not any cheaper'. I think the discussion suffered from translation issues on both sides. Cheaper as in less expensive, not lower quality. I can see how it could get confused and is a sensitive topic. It was just a one line statement and response. Not a big deal.

I just cant help but thinking only applying paste ,components , traces to one side, EDIT-mistaken on vias no vias thru the board, only inspect one side, has to be 'less expensive'

[Kilrah]

Well, you got a taste of how it is to deal with the language and usage barrier is - lots of that during the Taranis development as well, but it seems to have actually worked out pretty well in the end

And yes I've always been told that double sided component fitting was expensive due to the need of gluing components so that they don't fall off. But that was a few years ago, and I wouldn't be surprised if now the machines were capable of placing the components on one side, sloder, place the other side, solder again in no time. Haven't seen dots of glue for ages.
Same for double sided board manufacturing, barely anything is single-sided nowadays, so it must actually take them more time to adjust to it now

[rperkins]

Haven't seen dots of glue for ages.

Yea, They were phasing that out in 1988 at Wavetek RF products in Beech Grove Indiana. We populated the boards that had been etched elsewhere due to tighter environmental rules. In the mid '80s they even etched the boards on site. They went to a silkscreened solder paste setup instead of the epoxy glue. I have no idea what they do now, but whatever the process, I think 1 side is 'less expensive'.