Universal voltage-level translation

[[email protected]]

Hi !

I'm designing an universal JTAG programming/debugging dongle. I've
found it a bit problematic to find good logic level shifters for my
design. Basically I need output buffers interfacing 3.3V MCU with
"device under test" which may operate at 1.8V-5V.

My needs are as follows:

- all buffers are controlled by MCU with 3.3V I/O
- outputs can operate at any voltage between 1.8V and 5V, output Vcc is
taken from the "device under test"
- outputs have to be three-state kind. I need 3 signals with HiZ
controlled in common, and 2 signals with separate OEs.

Quick research led me to the following:

1. I can use dual-supply level translation from TI, series SN74LVC1T45,
SN74LVC2T45 etc. they have the right voltage ranges, unfrotunately no
three-state capabilities (only bigger ones, 8bits wide and wider).

2. Maxim's dual-supply translators, like MAX3390 - there are many
different input/output configurations so I could choose something for
me, however there are 2 requirements complicating the thing:
- one of the supply voltages (called VL) always need to be lower or
equal compared to another one (VCC) - I can eventually use 5V MCU I/O
- THREESTATE has to be driven from VL side I need it to be driven by
MCU, so from VCC side.

3. I've found NC7SZ125 from Fairchild. It's three-state buffer with VCC
range 1.65V - 5V, and overvoltage tolerant inputs - quite ok for me.
Still I need 5V MCU I/O to drive its inputs, if I want to support full
VCC range, but I can live with it (additional 3.3V -> 5V shifter would
do the trick). Does anyone know similar parts, with such a nice VCC
range and 5V-tolerant inputs ? Maybe more than one buffer per chip ?

Anyone here designed similar thing ? Any suggestions welcome..

BR,
Przemek

 

[Klaus Bahner]

do the trick). Does anyone know similar parts, with such a nice VCC

range and 5V-tolerant inputs ? Maybe more than one buffer per chip ?

Have you looked at TI's 74LVC2T45, which are dual rail 1.65 to 5.5V
bidirectional transceivers. Might be overkill for your application, but
the solved all level shifting problems I've encountered so far.

Klaus

 

[[email protected]]

Have you looked at TI's 74LVC2T45, which are dual rail 1.65 to 5.5V
bidirectional transceivers. Might be overkill for your application, but
the solved all level shifting problems I've encountered so far.

Yes, I even mentioned this part in my post - problem is that it lacks
OE input (HiZ mode).

BR,
Przemek

 

[Jim Thompson]

bidirectional transceivers

Have you already seen this note...

http://focus.ti.com/lit/an/scea035a/scea035a.pdf

...Jim Thompson

 

[Tim Shoppa]

I'm designing an universal JTAG programming/debugging dongle. I've
found it a bit problematic to find good logic level shifters for my
design. Basically I need output buffers interfacing 3.3V MCU with
"device under test" which may operate at 1.8V-5V.

My needs are as follows:

- all buffers are controlled by MCU with 3.3V I/O
- outputs can operate at any voltage between 1.8V and 5V, output Vcc is
taken from the "device under test"
- outputs have to be three-state kind. I need 3 signals with HiZ
controlled in common, and 2 signals with separate OEs.

Many commercial programmers meet similar needs with not digital
buffers, but with analog parts, e.g. op-amps (sometimes with push-pull
pairs on the output for extra current push or to explicitly tri-state
the output) and comparators rigged as schmitt triggers, and
digitally-programmed power supplies setting the rails. I think some TI
JTAG app notes show example circuits.

Do you have any specs for rise/fall time, capacitive load, ?

Tim.

 

[Przemyslaw Wegrzyn]

Jim Thompson wrote:

Have you already seen this note...

http://focus.ti.com/lit/an/scea035a/scea035a.pdf

I've just found it today, however I've already seen most of TI shifters
datasheets Still it's a good review of what TI offers.

Best Regards,
Przemek

 

[Przemyslaw Wegrzyn]

Many commercial programmers meet similar needs with not digital
buffers, but with analog parts, e.g. op-amps (sometimes with push-pull
pairs on the output for extra current push or to explicitly tri-state
the output) and comparators rigged as schmitt triggers, and

I've seen fast comparators used for inputs, already. Yet I've never seen
outputs done in analog way. Any schematics,articles,app notes welcome.

digitally-programmed power supplies setting the rails. I think some TI
JTAG app notes show example circuits.

As above, any pointers welcome. I'll search TI's site later today.

Do you have any specs for rise/fall time, capacitive load, ?

Not yet, really

BR,
Przemek

 

[Jim Thompson]

Jim Thompson wrote:



I've just found it today, however I've already seen most of TI shifters
datasheets Still it's a good review of what TI offers.

Best Regards,
Przemek

Personally, I just roll my own. Unfortunately they're not stand-alone
chips, they're part of a much larger ASIC. I think there's a hidden
example in one of my USB patents.

...Jim Thompson

 

[Tim Shoppa]

I've seen fast comparators used for inputs, already. Yet I've never seen
outputs done in analog way. Any schematics,articles,app notes welcome.


As above, any pointers welcome. I'll search TI's site later today.

I searched my desk but didn't see what I thought I remembered from TI's
app notes.

But if you look at Microchip's TB013 they show nice drivers for
delivering arbitrary programming voltages with op-amps and push-pull
drivers. A small change lets you tristate the drivers.

Tim.

 

[Przemyslaw Wegrzyn]

Tim Shoppa wrote:

But if you look at Microchip's TB013 they show nice drivers for
delivering arbitrary programming voltages with op-amps and push-pull
drivers. A small change lets you tristate the drivers.

I've finally found a little time to review it - the schematic there seems to
contain OP-AMP voltage followers with push-pull output stages, i.e. simple
current amps.

According to the app note, it's mainly needed due to capacitive loads on
VPP/RESET line. In my case I will not need such a high drive capabilities,
yet thanks for this pointer.

BR,
Przemek