Searching \ for '[PIC] Recommendations for a good gang programmer f' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/microchip/devprogs.htm?key=programmer
Search entire site for: 'Recommendations for a good gang programmer f'.

Exact match. Not showing close matches.
PICList Thread
'[PIC] Recommendations for a good gang programmer f'
2009\02\28@154914 by Vitaliy

flavicon
face
We're shopping for an industrial-strength gang programmer for the 24H
series. Preferably stand alone, and fast.

2009\02\28@172325 by olin piclist

face picon face
Vitaliy wrote:
> We're shopping for an industrial-strength gang programmer for the 24H
> series. Preferably stand alone, and fast.

Are you really sure you want a gang programmer, or just want to program a
bunch of PICs as fast as possible.  Gang programming has some inherent
problems, like separate verification of each PIC.

How about a bunch of USBProg or USBProg2 (http://www.embedinc.com/products)
running simultaneously?  Even after adding in the price of a powered USB
hub, you can probably do at least 10 PICs in parallel for less cost than a
true gang programmer, and you get the added advantage of separate readback
and error checking on each PIC.  This setup can also "gang" program PICs on
finished boards in circuit.



********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2009\02\28@191840 by Marcel

picon face
Vitaliy wrote:
> We're shopping for an industrial-strength gang programmer for the 24H
> series. Preferably stand alone, and fast.
>

You are in a position best suited to know what you need but your
emphasis on fast brings up questions like what package are your pics
going to be?  I have found that smd pics take longer to carefully insert
and remove from the socket and replace back in the tube they came out of
than the time it takes to program them - sometimes anyway.

A few now and then probably isn't worth the cost of a gang programmer.
If you are doing 100 or more, it might be worth looking at getting them
delivered to you already programmed.  This may not work if you need to
put different data in each.  For some '876 chips I get Digikey to
program them at something like 50 cents or something in 100 quantity.
They add a nice label with a legible program name as well.

I need to program about 50 of 8 pin soics once or twice a year so I do
it by my self.  Programming takes around 4 seconds; getting them out of
the tube, into the socket, out of the socket, back into the tube,
correctly oriented for the assembly house takes around 15 to 20 seconds.
 Having a gang unit won't speed this up.

Dip packages are better in this regard but not much.  Many years ago, I
did benefit from a gang unit that took around 5 minutes to burn 2708's.
 That definitely was worth plugging 8 parts into, hit "burn" and do
something else.


'[PIC] Recommendations for a good gang programmer f'
2009\03\01@094811 by olin piclist
face picon face
Marcel wrote:
> You are in a position best suited to know what you need but your
> emphasis on fast brings up questions like what package are your pics
> going to be?  I have found that smd pics take longer to carefully
> insert and remove from the socket and replace back in the tube they
> came out of than the time it takes to program them - sometimes anyway.

Good point.  This is another reason in-circuit programming in the built unit
is a good idea.  You have to build some sort of test fixture to test final
units and possibly calibrate and serialize them anyway.  The incremental
cost and time of adding programming to this step is usually minimal.  This
is likely cheaper than the cost of getting PICs programmed up front or the
the cost of the special handling to program them yourself, keeping track of
the programmed PICs with separate firmware as separate parts, and the cost
of scrapping them when the firmware changes every two weeks ;-)

If volumes are high and you need more throughput, set up multiple copies of
the same test station.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2009\03\01@132931 by Vitaliy

flavicon
face
Marcel wrote:
>> We're shopping for an industrial-strength gang programmer for the 24H
>> series. Preferably stand alone, and fast.
>
> You are in a position best suited to know what you need but your
> emphasis on fast brings up questions like what package are your pics
> going to be?  I have found that smd pics take longer to carefully insert
> and remove from the socket and replace back in the tube they came out of
> than the time it takes to program them - sometimes anyway.

I should have mentioned that the PICs will be programmed in-circuit.


> A few now and then probably isn't worth the cost of a gang programmer.
> If you are doing 100 or more, it might be worth looking at getting them
> delivered to you already programmed.  This may not work if you need to
> put different data in each.  For some '876 chips I get Digikey to
> program them at something like 50 cents or something in 100 quantity.
> They add a nice label with a legible program name as well.

Hm, interesting. What kind of labels do they use?

Vitaliy

2009\03\01@195824 by Marcel Duchamp

picon face
Vitaliy wrote:
>
>
>> A few now and then probably isn't worth the cost of a gang programmer.
>> If you are doing 100 or more, it might be worth looking at getting them
>> delivered to you already programmed.  This may not work if you need to
>> put different data in each.  For some '876 chips I get Digikey to
>> program them at something like 50 cents or something in 100 quantity.
>> They add a nice label with a legible program name as well.
>
> Hm, interesting. What kind of labels do they use?
>
> Vitaliy

Adhesive labels probably printed with an inkjet.  They cover nearly the
entire surface.

2009\03\01@201222 by Xiaofan Chen

face picon face
On Mon, Mar 2, 2009 at 2:27 AM, Vitaliy <spam_OUTspamTakeThisOuTspammaksimov.org> wrote:
> I should have mentioned that the PICs will be programmed in-circuit.

Typically programmed in-circuit is not what a gang programmer do.
Usually Gang programmers (BP, Data/IO, etc) do off-line programming.

>> A few now and then probably isn't worth the cost of a gang programmer.
>> If you are doing 100 or more, it might be worth looking at getting them
>> delivered to you already programmed.

Good suggestion.

Just take note that Microchip also provides such service (quick
turn programming). Last time, they only accepted large quantities
(>=25k?) which only adds about 5 US cents, lately they also
accept lower quantity.

Xiaofan

2009\03\01@203332 by Bob Blick

face
flavicon
face
Xiaofan Chen wrote:
> On Mon, Mar 2, 2009 at 2:27 AM, Vitaliy <.....spamKILLspamspam@spam@maksimov.org> wrote:
>> I should have mentioned that the PICs will be programmed in-circuit.
>
> Typically programmed in-circuit is not what a gang programmer do.
> Usually Gang programmers (BP, Data/IO, etc) do off-line programming.

I like the TI gang programmer for the MSP430. It is $199 and does
in-circuit just fine. Holds the program inside so you don't need a
computer. Hit the button and 8 green lights show up a few seconds later.
If you get a red light you have a bad board that wouldn't verify. Too
bad there isn't something like that for the PIC.

Lay your board out so one whole edge can go in a big edge connector.
Just put a small edge pattern for the programming signals. It works
great. If your board has lots of stuff near the edge, use a body grinder
and cut away the fixture except where the programming contacts are.

-Bob

2009\03\02@000052 by William \Chops\ Westfield

face picon face

On Mar 1, 2009, at 5:12 PM, Xiaofan Chen wrote:

>> I should have mentioned that the PICs will be programmed in-circuit.
>
> Typically programmed in-circuit is not what a gang programmer do.
> Usually Gang programmers (BP, Data/IO, etc) do off-line programming.

Huh.  True, and perhaps an indication of a hole in the market.  
Certainly I can imagine a production-floor gadget with a dozen  
tentacles coming out that you'd plug into your virgin products, push a  
button, and move on to the next.  By the time you got to the 12th, the  
first would be done, so you could pretty much produce boxes at the  
rate you could (carefully) insert and remove cables.  With a bit of  
customization, you could have the product report some self-test status  
back to the programmer, and get at least a first-effort idea of  
whether it was working.

On the other hand, I observe that many "gang programmers" seem to  
throw you into the "high priced PROFESSIONAL" equipment category, and  
you might find it easier, cheaper, and more flexible to simply buy  
several of a lower-cost individual programmers.  (for example: http://microcontrollershop.com/product_info.php?cPath=112_97&products_id=2910&osCsid=89e21f9961c2e264b9ce8a3cb19418a0
 is a 4-channel ICSP programmer for $999; compare to Olin's USBProg  
at $80 each, or even the ProProg at $299)

BillW

2009\03\02@023425 by Vitaliy

flavicon
face
Olin Lathrop wrote:
> If volumes are high and you need more throughput, set up multiple copies
> of
> the same test station.

A PC-based in-circuit programmer would be fine, except for one thing: they
have lower security.

For example, if Microchip's PM3 was fast enough, we could use superglue to
make the memory card non-ejectable, and maybe use a different connector for
USB. Not foolproof, but it would provide a deterrent to keep the person who
is programming the ICs, from copying the code.

With a PC-based programmer, we would have to put the binary in a location
accessible by the PC used for programming.

Do you have a solution in mind, that we haven't considered?

Vitaliy

2009\03\02@045324 by Alan B. Pearce

face picon face
> I should have mentioned that the PICs will be programmed in-circuit.

Sounds to me like you need a bunch of Pickit-3 units, which AIUI can hold a
copy of the destination code internally, for doing multiple programming like
this. Chain each one to the test bench so they cannot be borrowed overnight
to filch the code, and you should be OK.

2009\03\02@081739 by olin piclist

face picon face
Vitaliy wrote:
> A PC-based in-circuit programmer would be fine, except for one thing:
> they have lower security.
>
> For example, if Microchip's PM3 was fast enough, we could use
> superglue to make the memory card non-ejectable, and maybe use a
> different connector for USB. Not foolproof, but it would provide a
> deterrent to keep the person who is programming the ICs, from copying
> the code.
>
> With a PC-based programmer, we would have to put the binary in a
> location accessible by the PC used for programming.

You didn't explain this requirement originally.  I had assumed you were
asking for something to use on your own production line.  So you plan on
giving a high volume programming setup to customers or a remote
manufacturer?  A little more information on what you are really trying to
accomplish would help.

Anyone that is determined can monitor the programming signals going to the
PIC and get a dump of memory that way.  What level of copying effort are you
trying to protect against?

If you only want to prevent casual copying by not having the HEX file
available in direct form anywhere, it could be encrypted.  The programming
software would then decrypt it on the fly.  My software doesn't currently
have this capability, but it would not be hard to add.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2009\03\02@122751 by Vitaliy

flavicon
face
Olin Lathrop wrote:
>> A PC-based in-circuit programmer would be fine, except for one thing:
>> they have lower security.
>
> You didn't explain this requirement originally.

No, I did not. My bad.


> I had assumed you were
> asking for something to use on your own production line.  So you plan on
> giving a high volume programming setup to customers or a remote
> manufacturer?  A little more information on what you are really trying to
> accomplish would help.

This programmer will be used in-house.


> Anyone that is determined can monitor the programming signals going to the
> PIC and get a dump of memory that way.

You're absolutely right, and that's not the type of security we're after.


> What level of copying effort are you
> trying to protect against?

We just want to make stealing the code a bit more difficult than copying the
file to a thumb drive. What I described before would be adequate.


> If you only want to prevent casual copying by not having the HEX file
> available in direct form anywhere, it could be encrypted.  The programming
> software would then decrypt it on the fly.  My software doesn't currently
> have this capability, but it would not be hard to add.

The problem is, your software can be copied, together with the encrypted HEX
file.

Vitaliy

P.S. I know that this thread makes me look paranoid, and in fact I more
often than not I tend to be on the other end of the spectrum, especially
when it comes to people I work with. However, we had an instance of theft
last year (material goods, not IP) by a trusted employee, and I was
astonished at what he was willing to risk (a job he confessed he liked,
respect of his peers, clean criminal record) for what, in the end, is a
trivial sum (a small fraction of his salary). So now, while I still believe
that trust is the basis of any relationship (personal, business,
employer-employee), I think it is wise to reduce the temptation.

2009\03\02@125519 by olin piclist

face picon face
Vitaliy wrote:
>> What level of copying effort are you
>> trying to protect against?
>
> We just want to make stealing the code a bit more difficult than
> copying the file to a thumb drive. What I described before would be
> adequate.
>
>> If you only want to prevent casual copying by not having the HEX file
>> available in direct form anywhere, it could be encrypted.  The
>> programming software would then decrypt it on the fly.  My software
>> doesn't currently have this capability, but it would not be hard to
>> add.
>
> The problem is, your software can be copied, together with the
> encrypted HEX file.

True.  It all comes down to how clever a dirtbag you want to bother
protecting against.  There is nothing you can do about a really clever
dirtbag with enough resources.  All you can do is make it not worth the
trouble for some level of trouble and skill.

I was envisioning the decryption key be passed to PIC_PROG as a command line
option.  You could easily write a one-off canned program to run PIC_PROG
with the specific HEX file name and decryption key from within another
program you only put the binary of on the production machine.  This program
can also check a few particulars about the machine and refuse to run with
misleading error messages if things don't look right.  You can check the
processor ID, the serial number of the C drive, the MAC address of your NIC,
several obscure files having to be in the right place with the right
date/time stamp, etc.  Of course any of these can be spoofed, but the casual
dirtbag won't know they need to be spoofed and may not know how.  Simple
obfuscation is probably good enough for the situation you describe.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2009\03\02@134047 by Isaac Marino Bavaresco

flavicon
face
Olin Lathrop escreveu:
{Quote hidden}

We use an USB hard-lock. The key is in the hard-lock, along with some
counter or serial-number range.
Your program only knows how to talk to the hard-lock, and the hard-lock
won't allow to program more than a specified number of devices.

You can even put the motherboard serial number in the hard-lock to
prevent (or make harder to) the thief to program devices outside your
company.

Regards,

Isaac

__________________________________________________
Faça ligações para outros computadores com o novo Yahoo! Messenger
http://br.beta.messenger.yahoo.com/

2009\03\02@142828 by Vitaliy

flavicon
face
Isaac Marino Bavaresco wrote:
> We use an USB hard-lock. The key is in the hard-lock, along with some
counter or serial-number range.
Your program only knows how to talk to the hard-lock, and the hard-lock
won't allow to program more than a specified number of devices.

You can even put the motherboard serial number in the hard-lock to
prevent (or make harder to) the thief to program devices outside your
company. <-

What type of hard-lock is it, can you share the manufacturer/model number?
And this software you're talking about, is it custom?

Vitaliy

More... (looser matching)
- Last day of these posts
- In 2009 , 2010 only
- Today
- New search...