constant voltage or amperage

David F. Spencer dspencer at
Thu Sep 18 13:56:48 EST 1997

In article <342015DC.4894 at>, fordb at wrote:

> js wrote:
> > 
> > Hi!
> > If I can remember my phyiscs, then as the gel heats up during the run,
> > Resistance increases. 
> Unfortunately this is incorrect. Certainly, in solid conductors one will
> generally see positive temperature coefficients, which John may be
> remembering from his physics. But, for at least a couple of reasons, the
> resistance of electrophoresis gels will *drop* with increasing
> temperature (that is, there is a negative temperature coefficient).
> Amongst the reasons are the the increased thermal mobility of ions in a
> hot gel. Also the Arrhenius relation dictates that the degree of
> ionization usually rises in aqueous salt solutions at higher
> temperatures. David Spencer's suggestion that pore size may increase
> with temperature may also contribute.

As Bryan has said (and I mentioned in my post but apparently not clearly
enough) the behaviour of sequencing (i.e. denaturing polyacrylamide with
TBE) gels is not like simple resistors, but more like semiconductors.  And
there is no question that heat increases the pore size of polyacrylamide
gels because dyes and the DNA/RNA ladder do run faster in warm gels than
cold, under constant voltage.

However I did a simple experiment yesterday and have a better idea of the
main reason that TBE-buffered gels have major drops in resistance with
increasing temperature.  An important point that I should mention, because
few readers are likely to know this, is that the combination of Tris and
boric acid in water solution, particularly but not exclusively in equimolar
mixes, is quite unusual in that the conductivity of these solutions is 2 to
3 fold less at RT than solutions of the same concentration of say Tris-Cl
(8.0), sodium acetate, ammonium acetate, etc.  This makes TBE (which has a
pH at RT of about 8.3) a great buffer because it has the buffering capacity
of Tris-Cl (actually higher for alkaline shifts because of the first boric
acid pK) but its lower conductivity means that the rate of electrolytic
breakdown of the buffer is less than for other buffers.  What I determined
yesterday is that the conductivity of TBE increases fairly dramatically
with temperature such that at 65C, TBE (at 50mM-50mM) has about 50% higher
conductivity than at RT (ca. 25C).  This is almost certainly more than what
would be predicted from the Arrhenius relation (or thermal mobilities)
mentioned by Bryan, which I would assume (I don't remember the equation)
depend on a simple linear relation to temperature in Kelvin.  There is also
the fact that Tris has a very high temperature coefficient for pK (an
inverse one) but I can't immediately predict how that would contribute to
these observations.  Another peculiar feature of TBE is that the base,
Tris, has a pKa of 8.1 (RT) but the first pKa of boric acid is about 9.1.

> Personally, I prefer to regulate on constant "power" (watts). I would be
> interested in people's objections to this both from performance and
> safety perspectives. I always set the other controls (voltage and
> current) so that they will default into regulation if the parameters
> extend too far. But, I suspect that some newer digitally controlled
> apparatus may not allow this sort of parallel default regulation.
> -Bryan

There is certainly nothing "wrong" with running constant power (or the
override method you mention), and once you figure the parameters out,
running gels at constant power/wattage will give the shortest run time,
particularly for gels run warm or hot, without frying the gel and cracking
plates.  My opposition to this is purely a practical one because fully
regulated power supplies are several-fold more expensive than simple
unregulated ones (which at low loads produce not bad constant voltage) and
when I started doing nucleic acid (originally RNA) sequencing in '79 we
could buy several Dankar or Blaircraft unregulated power supplies for the
price of a single constant power type. At that time a constant power LKB
was about $4500 CDN, while the cheaper, simpler ones cost less than $1000. 
And although our lab has a couple of fully regulated power supplies I still
use a BioRad model 2000/200.

As far as safety is concerned, using constant power, or the overide method
you use, would be no more dangerous than simple constant voltage provided
the power supply is set up properly at run time so that the voltage upper
limit is locked in (limiting the current would be less effective because
the levels typically needed for running many sequencing gels could be
lethal under the right conditions).

Dave Spencer

David F. Spencer, PhD
Dept. Of Biochemistry
Dalhousie University
Halifax, Nova Scotia

dspencer at
dspencer at

More information about the Methods mailing list