gang programming ATTiny DIP

Go To Last Post
39 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm looking for a cheap gang programmer for a ATTiny25/45/85 DIP-8 chip. I've seen programmers at prices around 3000$ and that's way to expensive for programming about 100 IC's.

We can build our own stuff so is it possible to daisy chain the SPI programming interface or is there some document explaining how to gang program ATTiny IC's?

Since our design contains (at least) 2 Atmel processors, is there an easy way to program those via 1 connector instead of a dedicated connector for each IC.

Last Edited: Wed. Mar 27, 2019 - 03:26 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You could make your own gang programmer. I guess that 100 chips could be done in less time than you have designed and built the gang programmer. Even the soldering of 10 ZIF sockets will take longer.

I would be very surprised if your design could share the same MISO, MOSI, SCK lines on each AVR. So you would need separate ISP headers. You might share some pins.

If you think that you might ever do 1000s of Tinys, it is worth making your gang programmer. However I doubt that anyone would want 1000s of DIP chips.

Your gang programmer could be:
10 ZIF sockets with RST, MISO, SCK paralleled.
Each MOSI pin controlled via a 1:10 decoder and gate.

You insert all 10 chips. Press go. Each MOSI is enabled in turn for 10 programming sequences. This is 'consecutive' rather than 'gang' programming

An LED can indicate which chip is currently gated.

Completely untested. As I said, you could do 100 chips one at a time quite easily. At say 10 secs to program + 5 secs insert/remove per chip, 100 chips will take 25 minutes.

A commercial gang programmer is a lot more complex. You pay for complex electronics and small sales quantities.

If you want to truly gang program, you can use 10 AVRISP-2 + 10 ZIF. Your manual insert/remove times are going to be similar. The programming time goes from 100 secs to 10 secs. e.g. 100 chips in 10 minutes.

David.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jan_dc wrote:
We can build our own stuff so is it possible to daisy chain the SPI programming interface or is there some document explaining how to gang program ATTiny IC's?
You mean simultaneous ("gang") programming of batches of 100 mcus, within a few minutes? Wow, that's quite a bit of production volume!

While you can share MOSI and SCK (and probably RST too), you still have to read back MISO individually. I could visualise building such a programmer and it sounds to be a relatively simple task (electronics & software; the mechanical arrangement is a different issue ;-) ). However, I'd expect the chips' loading/unloading procedure to take much longer time than the programming itself, so unless you are planning to go for some substantial robotics, it might pay off more to simply use simply two commercial programmers, the operator unloading/loading one while the other is programming the chip.

jan_dc wrote:

Since our design contains (at least) 2 Atmel processors, is there an easy way to program those via 1 connector instead of a dedicated connector for each IC.
I'd go for separate SCK and all other signals shared; but of course that means they can't be programmed both at a time.

JW

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

What is the point in all this effort for just 100 parts? How long does each take to program? Let's say it is 1 minute plugged into a ZIF socket in an STK500. So you employ a student from the local job centre for $50/hour for 3 hours say and the job is done. That's a whole lot less than $3000 for a professiional gang programmer.

Or to put it another way, your $3,000 at $50/hour could buy 60 hours so the student could spend 36 minutes for each of the 100 AVR!

(though if you want "gang programming" ask the student if he can bring a long a few of his mates ;-))

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

100 IC is not a gang, it's just a tiny batch which can be programmed by a single person in a couple of hours.

Warning: Grumpy Old Chuff. Reading this post may severely damage your mental health.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

David, all our products are with DIP IC's (on sockets) and manually mounted on PCB's. You would be surprised how expensive SMD production is compared to manual production ;). Not every product is made in batches of 10.000's. As a small company it's in our best interest to spend as little time as possible doing monkey work. It's more interesting to spend time developing and building a programmer than spending the same time in programming IC's.

Anyway, we're just starting to use Atmel products. I've some other plans for the same IC and then we're talking about a batch of 1000 pieces. And then 25 minutes becomes over 4 hours. I don't think you want to do this job for 4 hours.

Ideally would be ISP of several processors via 1 connector. Fyi, we've got our own 'programming connector': RJ45.

Wek, SPI is suitable for daisy chaining: connect MISO to MOSI of next IC. That (should be) the beauty of SPI. : access a bunch of IC's with just 3 lines. I assume that this is not possible for ATTiny.

The IC's don't have to be programmed at the same time. They have to be programmed automatically with as little user attention as possible.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jan_dc wrote:
Wek, SPI is suitable for daisy chaining: connect MISO to MOSI of next IC. That (should be) the beauty of SPI. : access a bunch of IC's with just 3 lines.
Not all SPI devices are tolerant for this; in other words, not all SPI devices are implemented as transparent shift registers where the SS's trailing edge latches the incoming data. Not to mention the 4 possible combinations of CPOL/CPHA.

jan_dc wrote:
I assume that this is not possible for ATTiny.
While the AVR's serial programming protocol (ISP) resembles SPI, it is not SPI, neither it is a transparent shift register.

jan_dc wrote:
The IC's don't have to be programmed at the same time. They have to be programmed automatically with as little user attention as possible.
The AVR's ISP protocol is public and you can find it in every datasheet. There are tons of commercial and open source programmers, chose yourself.

Jan (too :-)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

This is why it is so important to give figures. 1000 is a lot different to 100. A Tiny85 is a lot different to a Tiny25. A single assemble + program ( + test) is probably quicker than programming chips before assembly.

I cannot understand why you would have two Tinys on one pcb. It would be simpler to have one Mega.

All the same, put in some realistic numbers and you will get some practical decisions. For a start, larger batches mean it is worth your distributor doing the programming.

No. I have never even thought of daisy-chaining ISP. There are certainly some NOP commands, so you could quite possibly propagate a whole lot of SPI commands down a chain. If you want to chain the two AVRs on each pcb you still restrict yourself with your pin usage.

I doubt that it would be very time-efficient. ISP programming by SPI is limited by the bus speed. The time to program each flash page is trivial.

Has anyone ever thought of daisy-chaining?
The biggest problem is getting several AVRs into 'programming mode'

David.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

david.prentice wrote:
This is why it is so important to give figures. 1000 is a lot different to 100. A Tiny85 is a lot different to a Tiny25. A single assemble + program ( + test) is probably quicker than programming chips before assembly.
And a student in Slovakia works for a lot less than in Belgium (or the UK for that matter).... ;-)

JW

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes my $50/hour was a huge over-estimate. The minimum wage here is something like £5/hour ($8/hour). I think that's common across Europe.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

wek wrote:
jan_dc wrote:
Wek, SPI is suitable for daisy chaining: connect MISO to MOSI of next IC. That (should be) the beauty of SPI. : access a bunch of IC's with just 3 lines.
Not all SPI devices are tolerant for this; in other words, not all SPI devices are implemented as transparent shift registers where the SS's trailing edge latches the incoming data. Not to mention the 4 possible combinations of CPOL/CPHA.

I know. We've got a bunch of the same AD's and those are not tranparent :(. Pitty because it would reduce pin count a lot AND the AD's are to be read out at the same time anyway.

wek wrote:
jan_dc wrote:
I assume that this is not possible for ATTiny.
While the AVR's serial programming protocol (ISP) resembles SPI, it is not SPI, neither it is a transparent shift register.

jan_dc wrote:
The IC's don't have to be programmed at the same time. They have to be programmed automatically with as little user attention as possible.
The AVR's ISP protocol is public and you can find it in every datasheet. There are tons of commercial and open source programmers, chose yourself.

Jan (too :-)

I did found a lot of programmers and building something for a single chip is easy. For multiple chip's it's very expensive. I hoped that there was a cheaper/easier way.

I somehow expected that Atmel already anticipated that multiple controllers/devices would be used in the same design.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hey, Cliff, Europe does not end at the eastern borders of Germany! :-)

I think I could get here a student for such undemanding, relatively comfortable indoors work for maybe E3/hr. The official minimum monthly pay is around E320 here, that's less than E2/hr. Most of the people here would be more than happy if they would earn £5/hour.

I am not complaining, just stating the facts.

JW

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

david.prentice wrote:
This is why it is so important to give figures. 1000 is a lot different to 100. A Tiny85 is a lot different to a Tiny25. A single assemble + program ( + test) is probably quicker than programming chips before assembly.

I cannot understand why you would have two Tinys on one pcb. It would be simpler to have one Mega.

It's a design consideration. The Tiny25 (or Tiny85) can be used a low freq clock source with 2 channels. Depending on the instrument we need up to 4 of those. It can also be used as a DAC or even double DAC. Yes I could use a Mega but those are too big. The 8-pin Tiny barely fits.

Yes, I do things for a reason ;) And I like to think ahead. I've made technical decisions 4 years ago which started to have an impact just now.

david.prentice wrote:

All the same, put in some realistic numbers and you will get some practical decisions. For a start, larger batches mean it is worth your distributor doing the programming.

No. I have never even thought of daisy-chaining ISP. There are certainly some NOP commands, so you could quite possibly propagate a whole lot of SPI commands down a chain. If you want to chain the two AVRs on each pcb you still restrict yourself with your pin usage.

I doubt that it would be very time-efficient. ISP programming by SPI is limited by the bus speed. The time to program each flash page is trivial.

Has anyone ever thought of daisy-chaining?
The biggest problem is getting several AVRs into 'programming mode'

David.

We don't have a fixed distributor for our parts. Secondly: we need to be able to change the functionality ourself. I can ask a distributor to program 1000 pieces and 3 months later I need to change it.

Programming speed is not the issue here. It's human interaction time. Ideally programming and calibration/testing is even done at once (working on that one). I don't care if it takes an hour or 10 minutes. That is just a mater of quantity: if it take an hour, I all arrange more to be done in parrallel. I just don't want people to wait for a machine doing something. If it takes 1 minute a piece. That 1 minute is wasted because you can't do anything else in the mean time.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

wek wrote:
Hey, Cliff, Europe does not end at the eastern borders of Germany! :-)

I think I could get here a student for such undemanding, relatively comfortable indoors work for maybe E3/hr. The official minimum monthly pay is around E320 here, that's less than E2/hr. Most of the people here would be more than happy if they would earn £5/hour.

I am not complaining, just stating the facts.

JW

I'm now wondering if you all work in a large company?

Here we have 2 engineers (and I'm 1 of them) and 4 people working in production. So on a yearly basis we've got not that much time to do as much as possible (and more important: having fun with it).

It also takes quite some time finding students, telling them what to do, do the administration,... And more important: it's not a structural solution. If I can find a good solution for my current problem it opens new design possibilities for us.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jan_dc wrote:
I somehow expected that Atmel already anticipated that multiple controllers/devices would be used in the same design.
They did not, and as I already told above, you could have multiple separately programmable devices on one board at the expense of one extra signal (separate SCK) per device.

JW

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

Hey, Cliff, Europe does not end at the eastern borders of Germany!

Is Slovakia in the EU? I'm pretty sure the minimum wage thing is across the EU. If Slovakia isn't in the EU then I guess it plans to be (though given the finances perhaps not?). If so it will have to implement the laws that the god's in Brussels dictate (as the rest of us have to).

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Cliff wrote:
Is Slovakia in the EU?
Since 2004.

http://en.wikipedia.org/wiki/Tem...

It's clickable. The middle "have all" group, bottom row, leftmost flag.

Cliff wrote:
I'm pretty sure the minimum wage thing is across the EU.
You are wrong. The span is more than one order of magnitude (compare Bulgaria/Romania with Luxemburg/France).

http://en.wikipedia.org/wiki/Lis...

http://assembly.coe.int/Document...

This is how our minimum wage evolved during the last years:
http://www.podnikam.sk/userfiles...

Of course these are only numbers.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Not sure this helps. I have merely multiplexed the reset and daisy chained the other ISP lines. Keeping the reset low on the chips not being programmed, on chip being programed reset follows ISP programmer signal.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

KitCarlson wrote:
Not sure this helps. I have merely multiplexed the reset and daisy chained the other ISP lines. Keeping the reset low on the chips not being programmed, on chip being programed reset follows ISP programmer signal.
This might get tricky. I wonder how this works for you. In fact, the ISP programmer after some initial wiggling simply keeps reset low all the time, so the chips "not being programmed" with reset held low tend to enter programming mode simultaneously. Been there...

I also don't understand why would a chip held in reset be MOSI->MISO transparent for the daisy chaininig.

That's why I recommended to separate (or mux, for that matter) SCK.

JW

Last Edited: Fri. Jan 13, 2012 - 02:16 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think we have discussed several methods.

1. minimum # pins for ISP header on pcb.
2. distributor programming.
3. make your own 'consecutive' gang programmer.
4. make your own parallel gang programmer
5. use several AVRISP-2

You do need to put some figures into the equation. A Tiny85 can be programmed and verified in about 500ms if this is what you want to do.

Your employee is going struggle to operate ZIFs or open tubes of DIP chips. She is going to be busy !!!

You need to design a good bench layout and ZIF jigs / sockets.

I would have thought that (1) is a good method if you can afford the pins. After all, I presume you want to test your assembled boards.

David.

Edit. Providing a chip is in RESET, MISO is 3-state until the chip gets AC53xxxx command. You can either gate off MOSI or SCK to prevent other chips listening.
Daisy chaining means you need all MISOs active.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

You would be surprised how expensive SMD production is compared to manual production

Yes, that would indeed certainly surprise me. Is there indeed a socket on each board? That is an expense; inserting the chips can lead to bent pins; possible oxidation over time, bigger boards (are mote expensive), ... -- all the reasons that you don't see bank(s) of socketed ICs much any more.

Our volumes are also modest; a production run sent to a board house is in three figures typically. All are programmed (often with board test code) via ISP using ATAVRISP(2). No complaints from the board houses. Connect the ISP header (or hold it in place making contact) and reach over and hit Program All on ChipBlaster. During the few seconds of ISP time, it is just enough to move the previous board to the "Done" box and stage the next from the "Raw" box.

I guess I just don't agree with your premises, starting with the first I quoted above.

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

I cannot understand why you would have two Tinys on one pcb. It would be simpler to have one Mega.
...

It's a design consideration. The Tiny25 (or Tiny85) can be used a low freq clock source with 2 channels. Depending on the instrument we need up to 4 of those. It can also be used as a DAC or even double DAC. Yes I could use a Mega but those are too big. The 8-pin Tiny barely fits.


OK, so you add system complexity and cost with n DIP Tiny25-family on the same board, yet are concerned with space?

TQFP32 -- 9mm x 9mm
DIP8 -- 9.25mm x 7.9mm (and several of those)
DIP8 socket -- 10mm x 10.16mm (and several of those)

Re the "training of your students"--heck, here in the US we even have salespeople doing ISP for upgrades and config loads with ISP. If they can be trained...

I'm out, I guess.

Lee

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

david.prentice wrote:
I think we have discussed several methods.

1. minimum # pins for ISP header on pcb.
2. distributor programming.
3. make your own 'consecutive' gang programmer.
4. make your own parallel gang programmer
5. use several AVRISP-2

You do need to put some figures into the equation. A Tiny85 can be programmed and verified in about 500ms if this is what you want to do.

Your employee is going struggle to operate ZIFs or open tubes of DIP chips. She is going to be busy !!!

You need to design a good bench layout and ZIF jigs / sockets.

I would have thought that (1) is a good method if you can afford the pins. After all, I presume you want to test your assembled boards.

David.

Edit. Providing a chip is in RESET, MISO is 3-state until the chip gets AC53xxxx command. You can either gate off MOSI or SCK to prevent other chips listening.
Daisy chaining means you need all MISOs active.

Yes, I think I'll have to go for a bunch of AVRISP's or something similar and hook each up to a ZIF socket and see if I can write some script for it. It's just wiring pins from AVRISP to ZIF socket. So it's going to be option 5).

Daisy-chain would be SOOO lovely :D. About 7 years ago I was in a project where we made a large system (with custom made chip's, FPGA's and other stuff) about everything was hooked together via JTAG daisy chain. With a single interface we could program and control the whole rack...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you insist of having one "standard" interface to program multiple chips on a board, you can program one chip and let it program the others.

I do something similar on a board with a "big" "first" (an ATMega2561) and two "small" "second" and "third" (ATMega8-s). The 'M8s' firmware here fits easily into the 'M2561, but it's not impossible even if the "first" is the same in size or smaller than the "others" - they would simply be programmed piecewise.

In your case, you'd need to signal the "first" back to the programmer when it's done, so it can be programmed by the firmware for the "next one", or by it's production firmware as the last step.

The programmer is then custom made of course, but that's not that complicated either.

JW

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

theusch wrote:
Quote:

You would be surprised how expensive SMD production is compared to manual production

Yes, that would indeed certainly surprise me. Is there indeed a socket on each board? That is an expense; inserting the chips can lead to bent pins; possible oxidation over time, bigger boards (are mote expensive), ... -- all the reasons that you don't see bank(s) of socketed ICs much any more.

Our volumes are also modest; a production run sent to a board house is in three figures typically. All are programmed (often with board test code) via ISP using ATAVRISP(2). No complaints from the board houses. Connect the ISP header (or hold it in place making contact) and reach over and hit Program All on ChipBlaster. During the few seconds of ISP time, it is just enough to move the previous board to the "Done" box and stage the next from the "Raw" box.

I guess I just don't agree with your premises, starting with the first I quoted above.

Now think about annual quantities of 50 pieces and even less. Think about repairability (we've products running for 30 years now). We even have our own soldering machine. Going to SMD would mean either invest in the machines or search for a 3-rd party do to the job. And more important: we can change fast. In worst case we'll have to dump some empty PCB's.

We've once asked for prices for 3-rd party production and those where very comparable to the cost we currently have.

At the moment I've got a few items (little modules) in the running for SMD assembly and annual quantities are 2000. This is interesting because on SMD I can use components that are not available in through hole. The cost is indeed lower. BUT: not because it's SMD, it's (barely) cheaper because I can get rid of other components.

Oxidation is not really an issue. Size of boards is a design issue, not a cost issue. Housings and customization of those are the highest cost. Second is component cost.

It's all about quantity and at the moment our volume is way to low to have a financial benefit over SMD.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

theusch wrote:
Quote:

I cannot understand why you would have two Tinys on one pcb. It would be simpler to have one Mega.
...

It's a design consideration. The Tiny25 (or Tiny85) can be used a low freq clock source with 2 channels. Depending on the instrument we need up to 4 of those. It can also be used as a DAC or even double DAC. Yes I could use a Mega but those are too big. The 8-pin Tiny barely fits.


OK, so you add system complexity and cost with n DIP Tiny25-family on the same board, yet are concerned with space?

TQFP32 -- 9mm x 9mm
DIP8 -- 9.25mm x 7.9mm (and several of those)
DIP8 socket -- 10mm x 10.16mm (and several of those)

Re the "training of your students"--heck, here in the US we even have salespeople doing ISP for upgrades and config loads with ISP. If they can be trained...

I'm out, I guess.

Lee

I don't think you know the design: I need 2 clocks with selectable frequency between 40Hz and 6kHz, the timers of the main ATMega processor are all used and you can't go for bit banging pins. Trace space is also limited.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Even without knowing your design details I can see an obvious mistake - wrong MCU selection.

Warning: Grumpy Old Chuff. Reading this post may severely damage your mental health.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

wek wrote:
If you insist of having one "standard" interface to program multiple chips on a board, you can program one chip and let it program the others.

I do something similar on a board with a "big" "first" (an ATMega2561) and two "small" "second" and "third" (ATMega8-s). The 'M8s' firmware here fits easily into the 'M2561, but it's not impossible even if the "first" is the same in size or smaller than the "others" - they would simply be programmed piecewise.

In your case, you'd need to signal the "first" back to the programmer when it's done, so it can be programmed by the firmware for the "next one", or by it's production firmware as the last step.

The programmer is then custom made of course, but that's not that complicated either.

JW

That sounds like a good idea. I've to look into the whole programming thing and see if can do something with it.

At least it would be useful for building a programmer.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

wek,
Perhaps it worked for me because the other chips were reset low and tri-state their pins. I never encountered a problem.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

MBedder wrote:
Even without knowing your design details I can see an obvious mistake - wrong MCU selection.

Ok, I would be very happy to hear other suggestions. I haven't found anything yet.

design rules:
* through hole components
* low volume
* as small as possible
* as little components as possible
* MCU interface
* 2 clock outputs with each seperate and selectable frequency
* accuracy is not important

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I repeat - do it all with a single thoroughly selected MCU. And forget about using through hole components - all you will get is an expensive, bulky and rapidly ageing product.

Warning: Grumpy Old Chuff. Reading this post may severely damage your mental health.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you do go down your chosen route.... consider this.I program my boards with a simple pogo pin adapter/pcb wired to a 6 pin plug that plugs into my AVR ISPII. It also has a power socket connector so I can power up the board with its own Vcc without having to make the board 'live'.There's no 6 pin header on the board, just 6 pads tracked to the relevant pins. Set everything up, press on the pins to the pads and hit program. Done in less than a second.PCBs are small and are in a 20 up snap out panel that has ben flow soldered.( yes, with DIP TINY25s.)It's FAST, can be done by monkeys and is field upgradable.I'd post a picture, but don't know how. If you pm me your email address, I'll fwd it to you.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

...like this you mean?
http://www.tag-connect.com/

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

MBedder wrote:
I repeat - do it all with a single thoroughly selected MCU. And forget about using through hole components - all you will get is an expensive, bulky and rapidly ageing product.

Nice to know that you now our production line and how we work. Please give me an estimate on how much our costs could be reduced. Suggestions for a good MCU for our designs would also be appreciated. And please be specific, I've no use for general replies.

Also I am very interested in how fast our products age if we should use SMD.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

To be honest, nobody knows your market, products, turnover, production system ...

I suspect that you would not want to broadcast them on a public website. However some private advice, or just a fresh pair of eyes can be very useful.

Yes. You can only make business decisions based on actual or projected figures.

David.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jan_dc wrote:
Please give me an estimate on how much our costs could be reduced. Suggestions for a good MCU for our designs would also be appreciated. And please be specific, I've no use for general replies
No problem. Normally I charge at least $10,000 for such analysis - please be ready to pay in advance.

Warning: Grumpy Old Chuff. Reading this post may severely damage your mental health.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

@MartinM57

Quote:
...like this you mean?
http://www.tag-connect.com/

Exactly, but I made my own, but it doesn't 'clip in'.
The other suggestion of a processor programming the chips is a possible solution, but adds cost and size.I have a mega48 that can be programmed via ISP and it(mega48) can also reprogram the on board 64Mbit SPI flash.The 48 polls a port bit on power up to detect if it should go into ext flash program mode.
That is only convenient because both parts are required in the system design- adigital audio playback device.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have made 8 furnaces for ATMEL AVR / MCS51 for $ 200. if you need more than 8 furnaces like 16, 24 I can make it cheaply

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Anang, if you're selling stuff, then keep in the Marketplace

:: Morten

 

(yes, I work for Microchip, yes, I do this in my spare time, now stop sending PMs)

 

The postings on this site are my own and do not represent Microchip’s positions, strategies, or opinions.

Topic locked