does your DC motor have a wound field? doubt its permanent magnet...if its wound field, is there a stabilizing shunt winding in the motor- saw one of those reverse wired before that caused all sorts of issues.
at 500 rpm on the spindle, is your motor going above base speed? if so there might be some field regulator issues. we had some antique ge valutrols on 850/2600 motors with 120 volt fields, in certain faults at max rpm(@ 2600 field was only around 12 volts) the fault would cause field current to return to full power immediately- well, 10x over field=10x over armature till it slows...2~3000 volts across the com would do heavy damage to the drives and the commutator- but this was big stuff...on a smaller drive, MOVs are about all they use to filter out spikes, but overvoltage from fast field rise at over base speed will cause high armature voltage bigtime.
quick comments on how non PM DC motors work, for those that never used these 'old' things-
max torque is continuous up to base speed, horsepower is constant from base to max rpm.
field voltage at or below base speed is full, armature voltage ramps up to increase rpm...at base speed or above, armature remains at full voltage, but field decreases to increase rpm. (complete field current loss if unchecked results in motor runaway- rpm is unlimited until the rotor grenades).
some larger motors had a stabilizing field winding(very heavy gage field winding) that is in series with armature solely to prevent runaway- they still use the small gage field wires for the bulk of below base speed field supply. if the stabilizing field is reverse wired, load decreases field, motor has crazy stability issues...
most dc motors have a direct coupled tach. *IF* your tach coupling is damaged, it *might* slip more in one direction than the other, *might* cause tach to overrun during fast braking, causing drive to think it needs to brake harder, repeat, pop. we had some old reliance REO-20 tachs with a 'rubber bladed' coupler that did this often...we chanced to a low backlash 'real' metal coupling to prevent it from recurring... careful with those dc wires too- a coworker at another plant lost some ribs, a lung, nearly his arm from reverse wiring a lathe drive tach- it ran away and the chuck exploded before the motor, runaway DC is scary stuff...on a mill its not as likely to explode as on a large lathe, but still, dont let the main switch be far out of reach when messing with non permanent magnet DC motors. opening the switch under load often blows the drive if its in regen, but better than grenading something.
most newer drives(after the late 80's) are pretty good at sensing tach divergence/runaway, but old stuff was scary.
some older dc drives used a NC contact across the armature leads for 'fixed' dynamic braking, the NC contact just put a huge wirewound resistor across the armature whenever the contactor opened...havent seen any like this in a long time, but might look for burnt contacts in the contactor...when stopping from under 500, does the contactor 'click' off before the spindle stops? if so it might have this hard DB hookup, and the resistor might be opening with higher heating, causing a dc spike...doubt it, just tossing ideas out there.
stupid thing that might work out if you cant drop the gain in the controller- a R-C network on the 0-10 volt line might be able to 'ramp down' the reference, and a parallel reverse diode to prevent most of the delay on accel...the .6 drop of the reverse diode will cause a slow response at the end of the accel, but might be a workaround...i'm assuming this machine is like in a home workshop, peak response not a huge concern? seems odd they have a accel pot but not decel if its a regen drive...if db there should be a huge discharge resistor...hmm... good luck