Once you have a few resistors on a schematic, it's tempting to do what you want using existing values. That reduces the BOM and the pick-and-place setup. There are far more possibilities if you are willing to compromise things some, or put resistors in series and parallel. This is a 4-page schematic with 9 different resistors. Doing this is an enormous PITA. https://www.dropbox.com/s/aozh6m6mmmik6lb/T502_Parts.jpg?raw=1 Saved a DAC reference chip! Actually, having the DAC voltage ratiometric on +15 is a benefit here. That opamp should be stable with the tantalum load and a 1u ceramic at the DAC, but I'll verify that. -- Father Brown's figure remained quite dark and still; but in that instant he had lost his head. His head was always most valuable when he had lost it.
juggling parts
Started by ●November 14, 2021
Reply by ●November 14, 20212021-11-14
On a sunny day (Sun, 14 Nov 2021 08:39:13 -0800) it happened jlarkin@highlandsniptechnology.com wrote in <54e2pgl0ir9ucjhtrcq35c0dj7pkadrgf8@4ax.com>:> >Once you have a few resistors on a schematic, it's tempting to do what >you want using existing values. That reduces the BOM and the >pick-and-place setup. There are far more possibilities if you are >willing to compromise things some, or put resistors in series and >parallel. This is a 4-page schematic with 9 different resistors. Doing >this is an enormous PITA. > >https://www.dropbox.com/s/aozh6m6mmmik6lb/T502_Parts.jpg?raw=1 > >Saved a DAC reference chip! Actually, having the DAC voltage >ratiometric on +15 is a benefit here. > >That opamp should be stable with the tantalum load and a 1u ceramic at >the DAC, but I'll verify that.Or a second LM317 for the +3V?
Reply by ●November 14, 20212021-11-14
On Sun, 14 Nov 2021 17:11:29 GMT, Jan Panteltje <pNaOnStPeAlMtje@yahoo.com> wrote:>On a sunny day (Sun, 14 Nov 2021 08:39:13 -0800) it happened >jlarkin@highlandsniptechnology.com wrote in ><54e2pgl0ir9ucjhtrcq35c0dj7pkadrgf8@4ax.com>: > >> >>Once you have a few resistors on a schematic, it's tempting to do what >>you want using existing values. That reduces the BOM and the >>pick-and-place setup. There are far more possibilities if you are >>willing to compromise things some, or put resistors in series and >>parallel. This is a 4-page schematic with 9 different resistors. Doing >>this is an enormous PITA. >> >>https://www.dropbox.com/s/aozh6m6mmmik6lb/T502_Parts.jpg?raw=1 >> >>Saved a DAC reference chip! Actually, having the DAC voltage >>ratiometric on +15 is a benefit here. >> >>That opamp should be stable with the tantalum load and a 1u ceramic at >>the DAC, but I'll verify that. > >Or a second LM317 for the +3V?Yuk! That would involve an entire new chain of value juggling. The resistor ratio needs to be 1.4:1. I can get 3.125 volts with three resistors. The +15 charges a timing ramp through resistors, and the DAC feeds some timing comparators, so it's nice if they are ratiometric. 317's are not very good references. Data sheets suggest that an LM317 can't make 3 volts from 5 volts, so I'd have to power it off +24. -- Father Brown's figure remained quite dark and still; but in that instant he had lost his head. His head was always most valuable when he had lost it.
Reply by ●November 14, 20212021-11-14
On a sunny day (Sun, 14 Nov 2021 09:39:03 -0800) it happened jlarkin@highlandsniptechnology.com wrote in <1ch2pgh4echute3tflmia84j7uosmor4rv@4ax.com>:>On Sun, 14 Nov 2021 17:11:29 GMT, Jan Panteltje ><pNaOnStPeAlMtje@yahoo.com> wrote: > >>On a sunny day (Sun, 14 Nov 2021 08:39:13 -0800) it happened >>jlarkin@highlandsniptechnology.com wrote in >><54e2pgl0ir9ucjhtrcq35c0dj7pkadrgf8@4ax.com>: >> >>> >>>Once you have a few resistors on a schematic, it's tempting to do what >>>you want using existing values. That reduces the BOM and the >>>pick-and-place setup. There are far more possibilities if you are >>>willing to compromise things some, or put resistors in series and >>>parallel. This is a 4-page schematic with 9 different resistors. Doing >>>this is an enormous PITA. >>> >>>https://www.dropbox.com/s/aozh6m6mmmik6lb/T502_Parts.jpg?raw=1 >>> >>>Saved a DAC reference chip! Actually, having the DAC voltage >>>ratiometric on +15 is a benefit here. >>> >>>That opamp should be stable with the tantalum load and a 1u ceramic at >>>the DAC, but I'll verify that. >> >>Or a second LM317 for the +3V? > >Yuk! That would involve an entire new chain of value juggling.Really?>The resistor ratio needs to be 1.4:1. I can get 3.125 volts with three >resistors.What I do not get is: you sell expensive things in low quantity and try to save on a trimpot? or waste board space with funny resistor combinations? I mean I use 3 resistors to get a LM317 to output some specific voltage all the time.. mostly different ones. Sure you can use 1 Ohm resistors to do most things, but then you need many,,,, 10 Ohm would work too here.>The +15 charges a timing ramp through resistors, and the DAC feeds >some timing comparators, so it's nice if they are ratiometric. 317's >are not very good references. > >Data sheets suggest that an LM317 can't make 3 volts from 5 volts, so >I'd have to power it off +24.Power it from the 15V you made with the first one.
Reply by ●November 14, 20212021-11-14
On Sun, 14 Nov 2021 17:59:06 GMT, Jan Panteltje <pNaOnStPeAlMtje@yahoo.com> wrote:>On a sunny day (Sun, 14 Nov 2021 09:39:03 -0800) it happened >jlarkin@highlandsniptechnology.com wrote in ><1ch2pgh4echute3tflmia84j7uosmor4rv@4ax.com>: > >>On Sun, 14 Nov 2021 17:11:29 GMT, Jan Panteltje >><pNaOnStPeAlMtje@yahoo.com> wrote: >> >>>On a sunny day (Sun, 14 Nov 2021 08:39:13 -0800) it happened >>>jlarkin@highlandsniptechnology.com wrote in >>><54e2pgl0ir9ucjhtrcq35c0dj7pkadrgf8@4ax.com>: >>> >>>> >>>>Once you have a few resistors on a schematic, it's tempting to do what >>>>you want using existing values. That reduces the BOM and the >>>>pick-and-place setup. There are far more possibilities if you are >>>>willing to compromise things some, or put resistors in series and >>>>parallel. This is a 4-page schematic with 9 different resistors. Doing >>>>this is an enormous PITA. >>>> >>>>https://www.dropbox.com/s/aozh6m6mmmik6lb/T502_Parts.jpg?raw=1 >>>> >>>>Saved a DAC reference chip! Actually, having the DAC voltage >>>>ratiometric on +15 is a benefit here. >>>> >>>>That opamp should be stable with the tantalum load and a 1u ceramic at >>>>the DAC, but I'll verify that. >>> >>>Or a second LM317 for the +3V? >> >>Yuk! That would involve an entire new chain of value juggling. > >Really? > > >>The resistor ratio needs to be 1.4:1. I can get 3.125 volts with three >>resistors. > >What I do not get is: you sell expensive things in low quantity and try to save on a trimpot?Trimpots are expensive and need manual turning. DACs are automated and can have cal factors stored in cal tables. I'm using 16-bit DACs and there are no 16-bit trimpots. Some of my customers absolutely forbid trimpots, which I think is extreme. Once in a while they make sense. They say that if there's a pot, their techs will turn it.>or waste board space with funny resistor combinations?I have space on this one for a few more 0603s. It's inefficient to load a reel onto the pick-and-place just for one resistor. The LM317 is huge.>I mean I use 3 resistors to get a LM317 to output some specific voltage all the time.. >mostly different ones. >Sure you can use 1 Ohm resistors to do most things, but then you need many,,,, >10 Ohm would work too here. > > >>The +15 charges a timing ramp through resistors, and the DAC feeds >>some timing comparators, so it's nice if they are ratiometric. 317's >>are not very good references. >> >>Data sheets suggest that an LM317 can't make 3 volts from 5 volts, so >>I'd have to power it off +24. > >Power it from the 15V you made with the first one. >OK, but the little opamp works better. Same number of parts, much smaller, more stable, uses available values. That's the game here. -- Father Brown's figure remained quite dark and still; but in that instant he had lost his head. His head was always most valuable when he had lost it.
Reply by ●November 14, 20212021-11-14
On a sunny day (Sun, 14 Nov 2021 10:27:20 -0800) it happened jlarkin@highlandsniptechnology.com wrote in <slk2pghav4cqdg31k20lnr0i2m63k5t6ur@4ax.com>:>On Sun, 14 Nov 2021 17:59:06 GMT, Jan Panteltje ><pNaOnStPeAlMtje@yahoo.com> wrote: > >>On a sunny day (Sun, 14 Nov 2021 09:39:03 -0800) it happened >>jlarkin@highlandsniptechnology.com wrote in >><1ch2pgh4echute3tflmia84j7uosmor4rv@4ax.com>: >> >>>On Sun, 14 Nov 2021 17:11:29 GMT, Jan Panteltje >>><pNaOnStPeAlMtje@yahoo.com> wrote: >>> >>>>On a sunny day (Sun, 14 Nov 2021 08:39:13 -0800) it happened >>>>jlarkin@highlandsniptechnology.com wrote in >>>><54e2pgl0ir9ucjhtrcq35c0dj7pkadrgf8@4ax.com>: >>>> >>>>> >>>>>Once you have a few resistors on a schematic, it's tempting to do what >>>>>you want using existing values. That reduces the BOM and the >>>>>pick-and-place setup. There are far more possibilities if you are >>>>>willing to compromise things some, or put resistors in series and >>>>>parallel. This is a 4-page schematic with 9 different resistors. Doing >>>>>this is an enormous PITA. >>>>> >>>>>https://www.dropbox.com/s/aozh6m6mmmik6lb/T502_Parts.jpg?raw=1 >>>>> >>>>>Saved a DAC reference chip! Actually, having the DAC voltage >>>>>ratiometric on +15 is a benefit here. >>>>> >>>>>That opamp should be stable with the tantalum load and a 1u ceramic at >>>>>the DAC, but I'll verify that. >>>> >>>>Or a second LM317 for the +3V? >>> >>>Yuk! That would involve an entire new chain of value juggling. >> >>Really? >> >> >>>The resistor ratio needs to be 1.4:1. I can get 3.125 volts with three >>>resistors. >> >>What I do not get is: you sell expensive things in low quantity and try to save on a trimpot? > >Trimpots are expensive and need manual turning. DACs are automated and >can have cal factors stored in cal tables. I'm using 16-bit DACs and >there are no 16-bit trimpots. > >Some of my customers absolutely forbid trimpots, which I think is >extreme. Once in a while they make sense. They say that if there's a >pot, their techs will turn it. > > >>or waste board space with funny resistor combinations? > >I have space on this one for a few more 0603s. It's inefficient to >load a reel onto the pick-and-place just for one resistor. > >The LM317 is huge. > >>I mean I use 3 resistors to get a LM317 to output some specific voltage all the time.. >>mostly different ones. >>Sure you can use 1 Ohm resistors to do most things, but then you need many,,,, >>10 Ohm would work too here. >> >> >>>The +15 charges a timing ramp through resistors, and the DAC feeds >>>some timing comparators, so it's nice if they are ratiometric. 317's >>>are not very good references. >>> >>>Data sheets suggest that an LM317 can't make 3 volts from 5 volts, so >>>I'd have to power it off +24. >> >>Power it from the 15V you made with the first one. >> > >OK, but the little opamp works better. Same number of parts, much >smaller, more stable, uses available values. That's the game here.All true As far as trimmers go, I remember the story of that guy who brought in a defective radio and said: "All those screws in those metal cans were loose, I fastened all of those." I have trimpots in use in many things, some have been on for decades.
Reply by ●November 14, 20212021-11-14
On Sun, 14 Nov 2021 19:59:43 GMT, Jan Panteltje <pNaOnStPeAlMtje@yahoo.com> wrote:>On a sunny day (Sun, 14 Nov 2021 10:27:20 -0800) it happened >jlarkin@highlandsniptechnology.com wrote in ><slk2pghav4cqdg31k20lnr0i2m63k5t6ur@4ax.com>: > >>On Sun, 14 Nov 2021 17:59:06 GMT, Jan Panteltje >><pNaOnStPeAlMtje@yahoo.com> wrote: >> >>>On a sunny day (Sun, 14 Nov 2021 09:39:03 -0800) it happened >>>jlarkin@highlandsniptechnology.com wrote in >>><1ch2pgh4echute3tflmia84j7uosmor4rv@4ax.com>: >>> >>>>On Sun, 14 Nov 2021 17:11:29 GMT, Jan Panteltje >>>><pNaOnStPeAlMtje@yahoo.com> wrote: >>>> >>>>>On a sunny day (Sun, 14 Nov 2021 08:39:13 -0800) it happened >>>>>jlarkin@highlandsniptechnology.com wrote in >>>>><54e2pgl0ir9ucjhtrcq35c0dj7pkadrgf8@4ax.com>: >>>>> >>>>>> >>>>>>Once you have a few resistors on a schematic, it's tempting to do what >>>>>>you want using existing values. That reduces the BOM and the >>>>>>pick-and-place setup. There are far more possibilities if you are >>>>>>willing to compromise things some, or put resistors in series and >>>>>>parallel. This is a 4-page schematic with 9 different resistors. Doing >>>>>>this is an enormous PITA. >>>>>> >>>>>>https://www.dropbox.com/s/aozh6m6mmmik6lb/T502_Parts.jpg?raw=1 >>>>>> >>>>>>Saved a DAC reference chip! Actually, having the DAC voltage >>>>>>ratiometric on +15 is a benefit here. >>>>>> >>>>>>That opamp should be stable with the tantalum load and a 1u ceramic at >>>>>>the DAC, but I'll verify that. >>>>> >>>>>Or a second LM317 for the +3V? >>>> >>>>Yuk! That would involve an entire new chain of value juggling. >>> >>>Really? >>> >>> >>>>The resistor ratio needs to be 1.4:1. I can get 3.125 volts with three >>>>resistors. >>> >>>What I do not get is: you sell expensive things in low quantity and try to save on a trimpot? >> >>Trimpots are expensive and need manual turning. DACs are automated and >>can have cal factors stored in cal tables. I'm using 16-bit DACs and >>there are no 16-bit trimpots. >> >>Some of my customers absolutely forbid trimpots, which I think is >>extreme. Once in a while they make sense. They say that if there's a >>pot, their techs will turn it. >> >> >>>or waste board space with funny resistor combinations? >> >>I have space on this one for a few more 0603s. It's inefficient to >>load a reel onto the pick-and-place just for one resistor. >> >>The LM317 is huge. >> >>>I mean I use 3 resistors to get a LM317 to output some specific voltage all the time.. >>>mostly different ones. >>>Sure you can use 1 Ohm resistors to do most things, but then you need many,,,, >>>10 Ohm would work too here. >>> >>> >>>>The +15 charges a timing ramp through resistors, and the DAC feeds >>>>some timing comparators, so it's nice if they are ratiometric. 317's >>>>are not very good references. >>>> >>>>Data sheets suggest that an LM317 can't make 3 volts from 5 volts, so >>>>I'd have to power it off +24. >>> >>>Power it from the 15V you made with the first one. >>> >> >>OK, but the little opamp works better. Same number of parts, much >>smaller, more stable, uses available values. That's the game here. > >All true >As far as trimmers go, I remember the story of that guy who brought in a defective >radio and said: "All those screws in those metal cans were loose, I fastened all of those." > >I have trimpots in use in many things, some have been on for decades.One good use for small trimpots is to set the gain in wideband - hundreds of MHz - things like photodiode amps. Especially when the product is not digital. Analog multipliers are expensive and relatively slow. PIN diodes are great but don't work at low frequencies or DC. A tiny DIP switch and resistors are OK if you don't mind coarse resolution. There are some super wideband many-bits digital RF attenuators, for around $100 each. Selected resistors are a nuisance. I think the wonderful Murata 2mm pots are extinct. They worked at 1 GHz or so. And were horrible to adjust. -- Father Brown's figure remained quite dark and still; but in that instant he had lost his head. His head was always most valuable when he had lost it.
Reply by ●November 14, 20212021-11-14
On 14/11/2021 18:27, jlarkin@highlandsniptechnology.com wrote:> On Sun, 14 Nov 2021 17:59:06 GMT, Jan Panteltje > <pNaOnStPeAlMtje@yahoo.com> wrote: > >> On a sunny day (Sun, 14 Nov 2021 09:39:03 -0800) it happened >> jlarkin@highlandsniptechnology.com wrote in >> <1ch2pgh4echute3tflmia84j7uosmor4rv@4ax.com>: >> >>> On Sun, 14 Nov 2021 17:11:29 GMT, Jan Panteltje >>> <pNaOnStPeAlMtje@yahoo.com> wrote: >>> >>>> On a sunny day (Sun, 14 Nov 2021 08:39:13 -0800) it happened >>>> jlarkin@highlandsniptechnology.com wrote in >>>> <54e2pgl0ir9ucjhtrcq35c0dj7pkadrgf8@4ax.com>: >>>> >>>>> >>>>> Once you have a few resistors on a schematic, it's tempting to do what >>>>> you want using existing values. That reduces the BOM and the >>>>> pick-and-place setup. There are far more possibilities if you are >>>>> willing to compromise things some, or put resistors in series and >>>>> parallel. This is a 4-page schematic with 9 different resistors. Doing >>>>> this is an enormous PITA. >>>>> >>>>> https://www.dropbox.com/s/aozh6m6mmmik6lb/T502_Parts.jpg?raw=1 >>>>> >>>>> Saved a DAC reference chip! Actually, having the DAC voltage >>>>> ratiometric on +15 is a benefit here. >>>>> >>>>> That opamp should be stable with the tantalum load and a 1u ceramic at >>>>> the DAC, but I'll verify that. >>>> >>>> Or a second LM317 for the +3V? >>> >>> Yuk! That would involve an entire new chain of value juggling. >> >> Really? >> >> >>> The resistor ratio needs to be 1.4:1. I can get 3.125 volts with three >>> resistors. >> >> What I do not get is: you sell expensive things in low quantity and try to save on a trimpot? > > Trimpots are expensive and need manual turning. DACs are automated and > can have cal factors stored in cal tables. I'm using 16-bit DACs and > there are no 16-bit trimpots. > > Some of my customers absolutely forbid trimpots, which I think is > extreme. Once in a while they make sense. They say that if there's a > pot, their techs will turn it. > > >> or waste board space with funny resistor combinations? > > I have space on this one for a few more 0603s. It's inefficient to > load a reel onto the pick-and-place just for one resistor. > > The LM317 is huge. > >> I mean I use 3 resistors to get a LM317 to output some specific voltage all the time.. >> mostly different ones. >> Sure you can use 1 Ohm resistors to do most things, but then you need many,,,, >> 10 Ohm would work too here. >> >> >>> The +15 charges a timing ramp through resistors, and the DAC feeds >>> some timing comparators, so it's nice if they are ratiometric. 317's >>> are not very good references. >>> >>> Data sheets suggest that an LM317 can't make 3 volts from 5 volts, so >>> I'd have to power it off +24. >> >> Power it from the 15V you made with the first one. >> > > OK, but the little opamp works better. Same number of parts, much > smaller, more stable, uses available values. That's the game here. > > > > >Pots are your enemy... -- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus
Reply by ●November 14, 20212021-11-14
On Sun, 14 Nov 2021 22:14:47 +0000, TTman <kraken.sankey@gmail.com> wrote:>On 14/11/2021 18:27, jlarkin@highlandsniptechnology.com wrote: >> On Sun, 14 Nov 2021 17:59:06 GMT, Jan Panteltje >> <pNaOnStPeAlMtje@yahoo.com> wrote: >> >>> On a sunny day (Sun, 14 Nov 2021 09:39:03 -0800) it happened >>> jlarkin@highlandsniptechnology.com wrote in >>> <1ch2pgh4echute3tflmia84j7uosmor4rv@4ax.com>: >>> >>>> On Sun, 14 Nov 2021 17:11:29 GMT, Jan Panteltje >>>> <pNaOnStPeAlMtje@yahoo.com> wrote: >>>> >>>>> On a sunny day (Sun, 14 Nov 2021 08:39:13 -0800) it happened >>>>> jlarkin@highlandsniptechnology.com wrote in >>>>> <54e2pgl0ir9ucjhtrcq35c0dj7pkadrgf8@4ax.com>: >>>>> >>>>>> >>>>>> Once you have a few resistors on a schematic, it's tempting to do what >>>>>> you want using existing values. That reduces the BOM and the >>>>>> pick-and-place setup. There are far more possibilities if you are >>>>>> willing to compromise things some, or put resistors in series and >>>>>> parallel. This is a 4-page schematic with 9 different resistors. Doing >>>>>> this is an enormous PITA. >>>>>> >>>>>> https://www.dropbox.com/s/aozh6m6mmmik6lb/T502_Parts.jpg?raw=1 >>>>>> >>>>>> Saved a DAC reference chip! Actually, having the DAC voltage >>>>>> ratiometric on +15 is a benefit here. >>>>>> >>>>>> That opamp should be stable with the tantalum load and a 1u ceramic at >>>>>> the DAC, but I'll verify that. >>>>> >>>>> Or a second LM317 for the +3V? >>>> >>>> Yuk! That would involve an entire new chain of value juggling. >>> >>> Really? >>> >>> >>>> The resistor ratio needs to be 1.4:1. I can get 3.125 volts with three >>>> resistors. >>> >>> What I do not get is: you sell expensive things in low quantity and try to save on a trimpot? >> >> Trimpots are expensive and need manual turning. DACs are automated and >> can have cal factors stored in cal tables. I'm using 16-bit DACs and >> there are no 16-bit trimpots. >> >> Some of my customers absolutely forbid trimpots, which I think is >> extreme. Once in a while they make sense. They say that if there's a >> pot, their techs will turn it. >> >> >>> or waste board space with funny resistor combinations? >> >> I have space on this one for a few more 0603s. It's inefficient to >> load a reel onto the pick-and-place just for one resistor. >> >> The LM317 is huge. >> >>> I mean I use 3 resistors to get a LM317 to output some specific voltage all the time.. >>> mostly different ones. >>> Sure you can use 1 Ohm resistors to do most things, but then you need many,,,, >>> 10 Ohm would work too here. >>> >>> >>>> The +15 charges a timing ramp through resistors, and the DAC feeds >>>> some timing comparators, so it's nice if they are ratiometric. 317's >>>> are not very good references. >>>> >>>> Data sheets suggest that an LM317 can't make 3 volts from 5 volts, so >>>> I'd have to power it off +24. >>> >>> Power it from the 15V you made with the first one. >>> >> >> OK, but the little opamp works better. Same number of parts, much >> smaller, more stable, uses available values. That's the game here. >> >> >> >> >> >Pots are your enemy...When I see a schematic full of trimpots, it's a sure bet that the author doesn't understand things. The Review Of Scientific Instruments is full of trimpots. And diffamps. And current mirrors. -- Father Brown's figure remained quite dark and still; but in that instant he had lost his head. His head was always most valuable when he had lost it.
Reply by ●November 14, 20212021-11-14
On Monday, November 15, 2021 at 9:22:55 AM UTC+11, jla...@highlandsniptechn= ology.com wrote:> On Sun, 14 Nov 2021 22:14:47 +0000, TTman <kraken...@gmail.com> wrote: > >On 14/11/2021 18:27, jla...@highlandsniptechnology.com wrote:=20 > >> On Sun, 14 Nov 2021 17:59:06 GMT, Jan Panteltje <pNaOnSt...@yahoo.com>=wrote:=20> >>> On a sunny day (Sun, 14 Nov 2021 09:39:03 -0800) it happened jla...@=highlandsniptechnology.com wrote in <1ch2pgh4echute3tf...@4ax.com>:=20> >>>> On Sun, 14 Nov 2021 17:11:29 GMT, Jan Panteltje <pNaOnSt...@yahoo.c=om> wrote:=20> >>>>> On a sunny day (Sun, 14 Nov 2021 08:39:13 -0800) it happened jla..=.@highlandsniptechnology.com wrote in <54e2pgl0ir9ucjhtr...@4ax.com>: \ <snip>> >>>>>> This is a 4-page schematic with 9 different resistors. Doing this =is an enormous PITA.=20> >>>>>>=20 > >>>>>> https://www.dropbox.com/s/aozh6m6mmmik6lb/T502_Parts.jpg?raw=3D1==20 It is a pity that the drop-box image doesn't show much of the schematic. <snip>> >>>> The resistor ratio needs to be 1.4:1. I can get 3.125 volts with thr=ee resistors.=20> >>>=20 > >>> What I do not get is: you sell expensive things in low quantity and t=ry to save on a trimpot?=20> >>=20 > >> Trimpots are expensive and need manual turning. DACs are automated and=can have cal factors stored in cal tables. I'm using 16-bit DACs and the= re are no 16-bit trimpots.=20 Actually there are, or at least there used to be - Vishay did a rather expe= nsive trim-pot with multiple brushes which was a lot easier to set precisel= y. I earned a few brownie points with final test by modifying a board which= had been a swine to set up by replacing the standard Bournes 18-mm 25 turn= trimmer with the Vishay drop in part. What we lost on the expensive part w= e more than saved on the setting up time.> >>=20 > >> Some of my customers absolutely forbid trimpots, which I think is ext=reme. Once in a while they make sense. They say that if there's a pot, the= ir techs will turn it. So do graduate students. Covering the slotted adjusting screw with a blob o= f epoxy after the pot has been set up can be a wise precaution. =20 <snip>> >Pots are your enemy... > > When I see a schematic full of trimpots, it's a sure bet that the author =doesn't understand things.=20 It isn't. The author is showing an understanding of things that John Larkin= doesn't have to worry about. =20> The Review Of Scientific Instruments is full of trimpots. And diffamps. =And current mirrors. The Review of Scientific Instruments published essentially one-off designs = to be used by physicists who more or less know what they are doing. The cir= cuits aren't designed for volume production. Giving graduate students a tri= m-pot they can twiddle - under supervision - while watching what happens ca= n help the education process no end, and it is lot less risky that swapping= soldered-in select-on-test resistors, particularly when you have an over-c= onfident graduate student doing the soldering. Differential amplifiers and current mirrors can be useful, and they can loo= k cute. Peer-reviewed science is performance art - the aim is to educate yo= ur readers, but you have to get their attention first, so a certain amount = of smoke and mirrors (even current mirrors) can be handy. I have published = in Rev.Sci. Instrum. (if only rude comments about some of the less impressi= ve circuits) and I do have a cited paper in Measurement Science and Technol= ogy so I do know more about this than John Larkin - but probably a whole lo= t less than Phil Hobbs or Winfield Hill. --=20 Bill Sloman, Sydney