Have you given your gauge a calibration "tap" with your finger? Analogue type gauges are quite inaccurate, they're not bad to begin with but the higher the reading (lower the pressure) the more inaccurate they get, in my experience they can "stick" and a calibration tap on the dial can get a completely different reading.
I've just done an experiment here at work, I have an altitude chamber with both an analogue and a digital gauge attached, I vacced it down as far as it would go, the digital gauge read 30mbar and the analogue read 120mbar. Giving the analogue gauge a tap on the glass brought this to 100mbar - better but still a long way from the digital gauge. I'd get a digital gauge if you can. Also is your gauge "barometrically compensated" if so that might be affecting the reading.
Our chamber has 11 different connections/joints along the way from the pump for various gauges, air admittance valves, joints etc and yet still gets down to 30mbar (equivalent to around 75000ft) yours apparently has 8, you could try smearing silicone bathroom sealant over the joints and letting it cure overnight, we have done this to our setup and it helps a lot.
Our normal test chamber only has 6 joints and we regularly test at a simulated altitude of 130000ft, but maybe we have a more powerful pump than you.