I have never tried this but I had the idea of building an 'enclosed set of shelves' above my workbench with a Lucite or similar door with thermometers built in and a way of 'gently' stirring the air inside while having a heat exchanger and the ability to 'cool' the air inside (since my shop is more likely to be hot and the equipment inside would generate heat). It would need power and other connections (Test Leads, Coax, LAN, HPIB, USB, RS232, etc.) that would be through some sort of 'sealed' access. It might take some work and it might be quite cumbersome but it sounds doable and would permit putting the appropriate instrument under test inside to do some simple 'environmental testing'. I wonder if such 'environmental test chambers' of a small size are commercially available? Joe -----Original Message----- From: volt-nuts-bounces@febo.com [mailto:volt-nuts-bounces@febo.com] On Behalf Of Fred Schneider Sent: Wednesday, August 22, 2012 2:23 AM To: Discussion of precise voltage measurement Subject: Re: [volt-nuts] HP3457 So this is normal, hmm, my shop can be over 30 degrees when it is a hot summer and 15 degrees in a Cold winter so keeping it powered will not help in my case ( i am just an amateur voltnut so climatize my room is no option. But The service manual describes a test were you run it at 15 degrees, then at 38 degrees and then adjust diodecurrent until you get the best tempco. That would imply the thing should not react much on a few degrees. Would it be an option to ad tempcontrol. The meter itself is a sort of uncontrolled oven. If I add a temp sensor and find a way to keep the temp steady the meter should stay more stable. In this case cooling would be the best option because it gets over 45 - 50 degrees inside the meter. So it can heats it self pretty well, The problem is temp follows the oudside temp, so if Measure inside temp at for instance 18 or 20 degrees room temp and cool it if the inside gets hotter then the inner temp at 20 degrees outside, than it must be more stable. Maybe a piezo element from a coolerbox inside (the piezo is inside, a coolrib and fan on theoutside. I experimented with a coolrib on the cabinet straight above the transformer, that was allready an improvent. Or some aluminum tubes through the cabinet. Open on one side, a box on the other side with a fan in it sucking the air through the pipes and cooling the inside when needed, or copper tubing where i can solder ribs on. If not enough I can pomp water through it and keep the water in an isolated box at 30 degrees using an aquarium heater if needed. That would cover 95% of the year. Stupid idea ? Fred PA4TIM Op 22 aug. 2012 om 03:32 heeft John Phillips <john.phillips0@gmail.com> het volgende geschreven: > To get better performance form you Solarton keep it a constant Hot > temperature all the time. Do not turn it off! Do not let the room > temperature fluctuate more than a few degrees. The more stale the > better. Cooling in most cases will induce error. If you can keep it at > 20C all the time go for it but if you are more stable at 28C then 28 > is better. > > The longer you run it the more stable it will get. We do not turn off > any of our best meters unless we are moving them or the power gos off. > > > On Tue, Aug 21, 2012 at 4:21 PM, Fred Schneider <pa4tim@gmail.com> > wrote: > >> I looked at the DC specs of the 3457 and thats the same as the >> solartron. I am looking sub uV like the solartron. The 3457 does not >> have that. So I think, regarding your reactions and the price, i'd >> better first take a look at the solartron. Modify it a bit ( >> cooling-ribs on the outside, one on top above the transformer and one >> on the back where the regulators should be ( i'm told, not looked >> yet) maybe a fan blowing over the ribs.) replace the keylock with a >> switch ( i have not a good key, it works but not smooth) then put >> banana busses instead of the back nr2 input connector on the front >> and replace the mainsswitch to the front. ( instead of climbing on a >> hair to reach the switch) >> >> I printed the service manual of the 7061 and i think I first do the >> tests ( psu, frequency, input current ect) i also noted my zero >> problem is because I shorted both inputs but i need to short the >> reference and front inputs. That could explain why zero often does >> not work ( error, zero to large or something lile that) >> >> I bought a Keithley 2000 last year, it was comparable to the Fluke >> 8046 but around 400 euro cheaper. I bought it straight from Keithley >> including calibration. Besides that I have a keithley 196 that was >> just calibrated when I bought it. ( and a 199) >> >> I was aiming for the holy grale of meters and a 3457 turns out to be >> a whole different beast. Someone I know bought a 3456 for 225 euro, >> fresh calibrated, that is 6,5 digits too. So i wondered why is a 3457 >> so much more expensive. >> >> But I better wait and save more money in the hope to find one day an >> affortable 3458. One should always keep something to dream about :-) >> >> Fred >> >> >> >> >> Fred PA4TIM >> >> Op 21 aug. 2012 om 22:06 heeft John Phillips >> <john.phillips0@gmail.com> het volgende geschreven: >> >>> you can put your 3457A into statistics mode and pick up +/- >>> 29,999,999 counts from the display. >>> +/- 2,999,999 counts is all you can get form the normal /live mode. >>> >>> On Tue, Aug 21, 2012 at 12:52 PM, J. L. Trantham <jltran@att.net> >>> wrote: >>> >>>> Fred, >>>> >>>> 700 euro seems a bit high, at least by US standards, for a 3457A. >>>> It >> is a >>>> good meter and I have one. Though no longer supported, Agilent >>>> will >> still >>>> calibrate one for $200.22, according to their website. >>>> >>>> You might want to consider a Fluke 8846A. I have seen them sell >>>> for >> $900 >>>> new. I don't have one but it was going to be my next addition >>>> until I >> got >>>> a >>>> 3458A. >>>> >>>> What is the issue with the 7061? >>>> >>>> Joe >>>> >>>> -----Original Message----- >>>> From: volt-nuts-bounces@febo.com >>>> [mailto:volt-nuts-bounces@febo.com] On Behalf Of Fred Schneider >>>> Sent: Tuesday, August 21, 2012 2:00 PM >>>> To: volt-nuts@febo.com >>>> Subject: [volt-nuts] HP3457 >>>> >>>> I have a solatron 7061 but i am not very impressed with it. Must I >>>> go looking for an alternative or is this 7061 when i gave it a good >> checkup / >>>> overhaul a good meter ? >>>> I am offered a 3457 for 700 euro and the owner, will calibrate it >>>> (he >> has >>>> several calibrtors, standards and a couple of 3458 meters. My >>>> intuition tells me to trusts him ( and that had never let me down >>>> before)) >>>> >>>> But he also told it is only 7,5 digit over hpib, and has 6,5 digit >> display. >>>> >>>> My Keithly 2000 is 6,5 digits and over RS232 i get two digits >>>> extra. I >> once >>>> tried that, but that took me a lot of time to get it more ore less >> working, >>>> I am not so into computers and afraid I do not get it to work. >>>> >>>> Is it really better as my ( one year old 2000 ?) and better as my >> Solartron >>>> can be ? >>>> I am afraid a 3458 is outside my budget. >>>> >>>> Fred >>>> >>>> >>>> >>>> Fred PA4TIM _______________________________________________ >>>> volt-nuts mailing list -- volt-nuts@febo.com To unsubscribe, go to >>>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>> and follow the instructions there. >>>> >>>> >>>> _______________________________________________ >>>> volt-nuts mailing list -- volt-nuts@febo.com >>>> To unsubscribe, go to >>>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>> and follow the instructions there. >>>> >>> >>> >>> >>> -- >>> John Phillips _______________________________________________ >>> volt-nuts mailing list -- volt-nuts@febo.com >>> To unsubscribe, go to >> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>> and follow the instructions there. >> >> _______________________________________________ >> volt-nuts mailing list -- volt-nuts@febo.com >> To unsubscribe, go to >> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >> and follow the instructions there. >> > > > > -- > John Phillips > _______________________________________________ > volt-nuts mailing list -- volt-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts > and follow the instructions there. _______________________________________________ volt-nuts mailing list -- volt-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts and follow the instructions there.

Fabio, The manul states the inner temp can go up to 75 degrees, i measured over 50 so extra heating it looks not logical to me. It would be for sure more easy as cooling. Most important will be to get the temp stabel. A closet like joe suggested would be an option. You could let the front come out, that will give some temp leakage but an improvement. The meter produces enough heat to warm the closet and if you keep that temp at for instance 35 degrees all you need is a tempsensor and a fan ( with adjustable speed) . As long as the room temp stays onder 35 degrees there would not be a problem. I think I will try that. Replcing switches ect is not an option, no room. But instead of the rackmount ears I will mount to alu profiles and mount the current inputs, Reference input and bananabusses instead of the second input on the back. I make a small hole in the side to lead the wirer out. On the other side I mount a profile with a powerswitch and two bnc panel troughputs and cables to the trigger bncs. I opened the meter and followed the checks of chapter 4. First problem, I coul not find if i have pcb 3 or 13 and 5 or 15 40348C 70610503X D so i think it is 3 11134F 70610505Y F so i think it is 5 B11091 70610508X B is the clockboard The meter has a date stamp 1987 so i think it is 3 and 5 The PROMs have other numbers as stated , they are toshiba TC57-256D-25 the others are 2516-45JL from texas instruments and SP1 is closed and SP2 is open ( as for type 2764) Allmost al solder bridges that should be open are soldered, some i can not ind and some are there but not in the manual. All voltages ar within specs. Tested it with a variac like the manual tells. So this is good. But then i measured the clock voltage TP402 That is a bit low 2.7V instead of 3V (+/- 0,2) so i should adjust coil L401 but i can not find it. This meter is ex-belgium army, so it looks like it has some mods or is a very early model. So this makes adjusting everything a bit tricky. Thanks for the offer but the key I have more or less works. It is a bit frubbling but calibrating is not something I do every day and most times i leave the key just in the cal position. Fred PA4TIM Op 22 aug. 2012 om 09:58 heeft Fabio Eboli <fabioeb@quipo.it> het volgende geschreven: > Hello Fred, > I'm certainly not expert here, but your idea of ovenizing > the 7061 made me think aloud :) > > So here are some random toughts. > My 7060 has no feet on underside and I noticed that the > unit, zeroed on 10V sitting flat on its bottom, drifted > badly if raised a pair of cm from the table, so the first > consideration I can do is that the 7060 must be calibrated > exactly in the same thermal equilibrium than in use. > The 7061 is a different beast but the internals are not too > much different from the 7060. > > The top and bottom covers are thick plastic covers, so they > isolate a little the unit from the outside world but do not > spread the heat around. The front and back panels are aluminium > so probably if there is temperature difference between outside > and inside, the gradient should be maximum near the panels. > > The infamous Fischer connectors (I hate them ;) should make > a very good job at maintaining no gradient between connections > of the leads, so thermocouple effects should be controlled well > by these connectors, also if there is a strong gradient on the > panel. Naturally this is just a guess from me, no mesurement > involved, this is your speciality :) > > The last consideration is that if I would try to heat the 7061 > I would try to avoid at any cost hot spots, so the heating should > be spread on a large surface. > As a first try I would place an aluminium plate on the underside > of the instrument, in contact with the outer side of the plastic > cover, and try to heat up this plate, for example bolting on or > gluing transistors on it in a mesh pattern. > As a second step I would try to isolate the heating plate from > the outside with foam. Later I would check the meter behaviour > and perhaps see what happens with another heat spreader-insulator > on top cover. > > The problem is how to properly test it, you should have a sort > of climatic chamber, in winter it's easier, just grab a pair > of big enoug cartoons, one into the other and heat up the air > between them :) > > Ok this is enough fantasy at work :) > > As for the key, I can duplicate mine and send it to you, > if you want to try another key. > > Fabio. > > Fred Schneider <pa4tim@gmail.com> ha scritto: > >> So this is normal, hmm, my shop can be over 30 degrees when it is a hot summer and 15 degrees in a Cold winter so keeping it powered will not help in my case ( i am just an amateur voltnut so climatize my room is no option. >> >> But The service manual describes a test were you run it at 15 degrees, then at 38 degrees and then adjust diodecurrent until you get the best tempco. That would imply the thing should not react much on a few degrees. >> >> Would it be an option to ad tempcontrol. The meter itself is a sort of uncontrolled oven. If I add a temp sensor and find a way to keep the temp steady the meter should stay more stable. In this case cooling would be the best option because it gets over 45 - 50 degrees inside the meter. So it can heats it self pretty well, The problem is temp follows the oudside temp, so if Measure inside temp at for instance 18 or 20 degrees room temp and cool it if the inside gets hotter then the inner temp at 20 degrees outside, than it must be more stable. >> >> Maybe a piezo element from a coolerbox inside (the piezo is inside, a coolrib and fan on theoutside. I experimented with a coolrib on the cabinet straight above the transformer, that was allready an improvent. Or some aluminum tubes through the cabinet. Open on one side, a box on the other side with a fan in it sucking the air through the pipes and cooling the inside when needed, or copper tubing where i can solder ribs on. If not enough I can pomp water through it and keep the water in an isolated box at 30 degrees using an aquarium heater if needed. That would cover 95% of the year. >> >> Stupid idea ? >> >> >> Fred PA4TIM >> >> Op 22 aug. 2012 om 03:32 heeft John Phillips <john.phillips0@gmail.com> het volgende geschreven: >> >>> To get better performance form you Solarton keep it a constant Hot >>> temperature all the time. Do not turn it off! Do not let the room >>> temperature fluctuate more than a few degrees. The more stale the better. >>> Cooling in most cases will induce error. If you can keep it at 20C all the >>> time go for it but if you are more stable at 28C then 28 is better. >>> >>> The longer you run it the more stable it will get. We do not turn off any >>> of our best meters unless we are moving them or the power gos off. >>> >>> >>> On Tue, Aug 21, 2012 at 4:21 PM, Fred Schneider <pa4tim@gmail.com> wrote: >>> >>>> I looked at the DC specs of the 3457 and thats the same as the solartron. >>>> I am looking sub uV like the solartron. The 3457 does not have that. So I >>>> think, regarding your reactions and the price, i'd better first take a look >>>> at the solartron. Modify it a bit ( cooling-ribs on the outside, one on top >>>> above the transformer and one on the back where the regulators should be ( >>>> i'm told, not looked yet) maybe a fan blowing over the ribs.) replace the >>>> keylock with a switch ( i have not a good key, it works but not smooth) >>>> then put banana busses instead of the back nr2 input connector on the front >>>> and replace the mainsswitch to the front. ( instead of climbing on a hair >>>> to reach the switch) >>>> >>>> I printed the service manual of the 7061 and i think I first do the tests >>>> ( psu, frequency, input current ect) i also noted my zero problem is >>>> because I shorted both inputs but i need to short the reference and front >>>> inputs. That could explain why zero often does not work ( error, zero to >>>> large or something lile that) >>>> >>>> I bought a Keithley 2000 last year, it was comparable to the Fluke 8046 >>>> but around 400 euro cheaper. I bought it straight from Keithley including >>>> calibration. Besides that I have a keithley 196 that was just calibrated >>>> when I bought it. ( and a 199) >>>> >>>> I was aiming for the holy grale of meters and a 3457 turns out to be a >>>> whole different beast. Someone I know bought a 3456 for 225 euro, fresh >>>> calibrated, that is 6,5 digits too. So i wondered why is a 3457 so much >>>> more expensive. >>>> >>>> But I better wait and save more money in the hope to find one day an >>>> affortable 3458. >>>> One should always keep something to dream about :-) >>>> >>>> Fred >>>> >>>> >>>> >>>> >>>> Fred PA4TIM >>>> >>>> Op 21 aug. 2012 om 22:06 heeft John Phillips <john.phillips0@gmail.com> >>>> het volgende geschreven: >>>> >>>>> you can put your 3457A into statistics mode and pick up +/- 29,999,999 >>>>> counts from the display. >>>>> +/- 2,999,999 counts is all you can get form the normal /live mode. >>>>> >>>>> On Tue, Aug 21, 2012 at 12:52 PM, J. L. Trantham <jltran@att.net> wrote: >>>>> >>>>>> Fred, >>>>>> >>>>>> 700 euro seems a bit high, at least by US standards, for a 3457A. It >>>> is a >>>>>> good meter and I have one. Though no longer supported, Agilent will >>>> still >>>>>> calibrate one for $200.22, according to their website. >>>>>> >>>>>> You might want to consider a Fluke 8846A. I have seen them sell for >>>> $900 >>>>>> new. I don't have one but it was going to be my next addition until I >>>> got >>>>>> a >>>>>> 3458A. >>>>>> >>>>>> What is the issue with the 7061? >>>>>> >>>>>> Joe >>>>>> >>>>>> -----Original Message----- >>>>>> From: volt-nuts-bounces@febo.com [mailto:volt-nuts-bounces@febo.com] On >>>>>> Behalf Of Fred Schneider >>>>>> Sent: Tuesday, August 21, 2012 2:00 PM >>>>>> To: volt-nuts@febo.com >>>>>> Subject: [volt-nuts] HP3457 >>>>>> >>>>>> I have a solatron 7061 but i am not very impressed with it. Must I go >>>>>> looking for an alternative or is this 7061 when i gave it a good >>>> checkup / >>>>>> overhaul a good meter ? >>>>>> I am offered a 3457 for 700 euro and the owner, will calibrate it (he >>>> has >>>>>> several calibrtors, standards and a couple of 3458 meters. My intuition >>>>>> tells me to trusts him ( and that had never let me down before)) >>>>>> >>>>>> But he also told it is only 7,5 digit over hpib, and has 6,5 digit >>>> display. >>>>>> >>>>>> My Keithly 2000 is 6,5 digits and over RS232 i get two digits extra. I >>>> once >>>>>> tried that, but that took me a lot of time to get it more ore less >>>> working, >>>>>> I am not so into computers and afraid I do not get it to work. >>>>>> >>>>>> Is it really better as my ( one year old 2000 ?) and better as my >>>> Solartron >>>>>> can be ? >>>>>> I am afraid a 3458 is outside my budget. >>>>>> >>>>>> Fred >>>>>> >>>>>> >>>>>> >>>>>> Fred PA4TIM >>>>>> _______________________________________________ >>>>>> volt-nuts mailing list -- volt-nuts@febo.com To unsubscribe, go to >>>>>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>>>> and follow the instructions there. >>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> volt-nuts mailing list -- volt-nuts@febo.com >>>>>> To unsubscribe, go to >>>>>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>>>> and follow the instructions there. >>>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> John Phillips >>>>> _______________________________________________ >>>>> volt-nuts mailing list -- volt-nuts@febo.com >>>>> To unsubscribe, go to >>>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>>> and follow the instructions there. >>>> >>>> _______________________________________________ >>>> volt-nuts mailing list -- volt-nuts@febo.com >>>> To unsubscribe, go to >>>> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>>> and follow the instructions there. >>>> >>> >>> >>> >>> -- >>> John Phillips >>> _______________________________________________ >>> volt-nuts mailing list -- volt-nuts@febo.com >>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >>> and follow the instructions there. >> >> _______________________________________________ >> volt-nuts mailing list -- volt-nuts@febo.com >> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts >> and follow the instructions there. >> > > > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. > > > > _______________________________________________ > volt-nuts mailing list -- volt-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts > and follow the instructions there.

The 3457a is younger, has added functions, such as amps, and better long term accuracy; the 3456a is just a DVM. The previously posted link has a large comparative database: http://electronicsandbooks.com/eab1/kb/k0zfca/i0zfca7_.htm 3457a tends to be have more unobtanium, compared to the 3456a. For 700 euros for a calibrated 3457a, its better to look at new 6.5 digit DMMs such as the 34410a which I see farnell lists for 850 euros. http://uk.farnell.com/agilent-technologies/34410a/multimeter-bench-61-2-digit/dp/1440376 or the cheaper 34401a or equivalents from other makers, Fluke, Keithley etc.. again, electronicsandbooks has a wonderful comparative listing. At 07:21 PM 8/21/2012, Fred Schneider wrote: >I was aiming for the holy grale of meters and a 3457 turns out to be >a whole different beast. Someone I know bought a 3456 for 225 euro, >fresh calibrated, that is 6,5 digits too. So i wondered why is a >3457 so much more expensive. Sincerely, Marv Gozum Philadelphia, PA