25V H-bridge for <3V logic?

Thanks, interesting device. However, it needs a 10V supply which I'd have to generate since there will only be

Are you ramping the supplies on that from zero and back to zero?

Just in case others need this as well: The Zetex ZHB6790 looketh pretty cool:

Well, that plus a few granules for the high side drive. Nice min hfe ratings, too. The $1 price tag doesn't spread happiness here but there has got to be others or maybe some "haggling room" for pricing.

No I don't. I was aware of your requirements and I didn't think that the 358 meets them. I was just pointing out that a dual opamp can make a nice cheap slow H-bridge.

robert

That's easy for opamps but when you ramp the supplies to below spec some of them go ballistic. Problem is, this behavior can vary between manufacturers for the same part number.

I am going to use BJTs and FETs again, as usual. When looking for arrays today I was appalled. They want almost $1 for the ZHB6790 bridge. No way. So it'll be discrete. Again. Just this time SOT416, SOT523 and stuff like that. SOT23 looks like a huge rock in comparison. Ok in production but absolutely no fun when prototyping. At least no with the typical eyesight of a guy who is over the hill.

When you say "prorotyping" with this kind of packages, do you mean you're actually sitting down yorself and hand-solder the stuff? Or is it just about debugging machine-assembled prototypes?

Myself, I've been able to avoid SMD altogether to this day (I'm only doing low-volume "in house" stuff), but right now I'm space-constrained enough to have to do a SMD design with components on both sides of the PCB. Components are not smaller than 0805, SOT23 and SO8 so I hope to be able to pull it off without specialized SMD tools using only a regular fine-tipped soldering iron and normal tweezers. What worries me most isn't so much the assembly than possible rework.

robert

Even the first units are always professionally done but there is always a chance for cuts and jumpers. Mistakes, last minute spec changes, placement programming errors, the usual. Also, there is a lot of verification measurement. When there is 0.25mm between the pins one little itch can cause the probe to slip and the whole thing to fry. Some of the discrete parts look more like dandruff than electronics.

Ah, 0805, the good old days :-)

Here, it's mostly 0402 and now going to 0201. SOT23 is also almost a thing of the past for me.

So what /is/ all this super-small, super-cheap, high-volume stuff you're designing?

robert

On 18 Jan 2007 12:30:45 GMT, Robert Latest wrote in Msg.

Oh, just noticed you answered my question in another thread. Thanks. r.

It's actually more than that. Sometimes clients, especially new ones, retain me to lower the cost on one of their products. When swooping down into the schematic usually the first thing I do is look for fancy programmables. CPLD, FPGA, uC, but also those "we-do-it-all" panacea chip. A few price checks, a painful "ouch", and out they go. Back to analog can mean a substantial increase in component count and the only way to squeeze that into the existing form factor and enclosure is to drop smaller SMT sizes. Even if you just replace a FET by a cheaper BJT that adds at least one resistor. The first reactions in the design review are usually gasping and scoffing but the smiles come back when the burdened BOM cost total pops up. At least the financial guys are smiling :-)

with those size you shouldn't have any problems with a normal soldering iron and tweezers. 0603 is just as easy and possibly easier to get I don't think 0805 is used much anymore.

I'd say rework is easier with smd. exept for bgas and such ofcourse...

having two soldering irons helps too

-Lasse

That, and very small small diameter solder. Here in the lab the favorite is Kester no-clean 15-mil (-8806).

"Joerg" schrieb im Newsbeitrag news:FxNrh.2392$ snipped-for-privacy@newssvr11.news.prodigy.net... | Robert Latest wrote: | > On 18 Jan 2007 12:30:45 GMT, | > Robert Latest wrote | > in Msg. | >

| >>So what /is/ all this super-small, super-cheap, high-volume stuff you're | >>designing? | >

| >

| > Oh, just noticed you answered my question in another thread. Thanks. r. | >

| | It's actually more than that. Sometimes clients, especially new ones, | retain me to lower the cost on one of their products. When swooping down | into the schematic usually the first thing I do is look for fancy | programmables. CPLD, FPGA, uC, but also those "we-do-it-all" panacea | chip. A few price checks, a painful "ouch", and out they go. Back to | analog can mean a substantial increase in component count and the only | way to squeeze that into the existing form factor and enclosure is to | drop smaller SMT sizes. Even if you just replace a FET by a cheaper BJT | that adds at least one resistor. The first reactions in the design | review are usually gasping and scoffing but the smiles come back when | the burdened BOM cost total pops up. At least the financial guys are | smiling :-)

Going smaller is beautiful but I wonder why the FPGA stuff have so many pins today. Even 144 is "not much". Something like a paradoxum to me. Most of the FPGA pins typically does lower than DC - almost "nothing".

- Henry

--

"Robert Latest" schrieb im Newsbeitrag news: snipped-for-privacy@mid.individual.net... | Joerg wrote: | | > I am going to use BJTs and FETs again, as usual. When looking for arrays | > today I was appalled. They want almost $1 for the ZHB6790 bridge. No | > way. So it'll be discrete. Again. Just this time SOT416, SOT523 and | > stuff like that. SOT23 looks like a huge rock in comparison. Ok in | > production but absolutely no fun when prototyping. At least no with the | > typical eyesight of a guy who is over the hill.

Nice little chip. Now I can recall it as I see it. A good alternative to the ICL7667.

Sometimes it is difficult to recover all the nice silicone we can use.

Thanks Joerg!

- Henry

--

Why not ZDT6790 ? Much cheaper!

- Henry

--

Yeah but still around 40c. Nah, I've already done it with four discrete which costs a lot less.

"Joerg" schrieb im Newsbeitrag news:wdbsh.2596$ snipped-for-privacy@newssvr11.news.prodigy.net... | Henry Kiefer wrote: | | > Why not ZDT6790 ? | > Much cheaper! | >

| | Yeah but still around 40c. Nah, I've already done it with four discrete | which costs a lot less.

I dream of the day having a cheaper design than Joerg.

- Henry

--

news:wdbsh.2596$ snipped-for-privacy@newssvr11.news.prodigy.net...

The best attempt in that direction I saw in Korea. They had a wall full of long tractor paper printouts. They changed these out several times a week and I thought it was stock quotes. When taking a close look it was all component prices.

"Joerg" schrieb im Newsbeitrag news:Seesh.62716$ snipped-for-privacy@newssvr13.news.prodigy.net... | Henry Kiefer wrote: | | > "Joerg" schrieb im Newsbeitrag news:wdbsh.2596$ snipped-for-privacy@newssvr11.news.prodigy.net... | > | Henry Kiefer wrote: | > | | > | > Why not ZDT6790 ? | > | > Much cheaper! | > | >

| > | | > | Yeah but still around 40c. Nah, I've already done it with four discrete | > | which costs a lot less. | >

| > I dream of the day having a cheaper design than Joerg. | >

| | The best attempt in that direction I saw in Korea. They had a wall full | of long tractor paper printouts. They changed these out several times a | week and I thought it was stock quotes. When taking a close look it was | all component prices.

Social engineering is maybe a better solution to have good prices.

Interesting - Henry

--