03-02-2020, 04:21 PM
If I'd known about that pythagoras circuit in 1980/81 it might have saved me a great deal of grief. Though strictly I would have needed 4 quadrant operation so I would have had to rectify the inputs first. Might not have been easy to make with discretes or even arrays like the CA3046.
I did actually read one of Gilbert's papers back then and did some interesting experiments on improving the accuracy of commerically available Gilbert cell chips such as the 1496 and TBA673: https://datasheetspdf.com/parts/TBA673.pdf?id=1251598 I was using them for video gain control in vision mixer designs and wanted their operation to be totally invisible. This wasn't easy. How trivial it is now in digits. Though back then it was hard in digits too, but for different reasons. The need for a lot of logic running at what was then high speed (13.5MHz). I'll write a note about the TDC1007 video ADC in a moment.
I did actually read one of Gilbert's papers back then and did some interesting experiments on improving the accuracy of commerically available Gilbert cell chips such as the 1496 and TBA673: https://datasheetspdf.com/parts/TBA673.pdf?id=1251598 I was using them for video gain control in vision mixer designs and wanted their operation to be totally invisible. This wasn't easy. How trivial it is now in digits. Though back then it was hard in digits too, but for different reasons. The need for a lot of logic running at what was then high speed (13.5MHz). I'll write a note about the TDC1007 video ADC in a moment.
www.borinsky.co.uk Jeffrey Borinsky www.becg.tv
