For those interested, Tenma 72-6644 is actually produced by METEX and is MXG-9810. It is also marketed under different names, for instance, Peaktech 4085
There are several manuals which include the bill of materials, PCBs and schematics
(
https://elektrotanya.com/cgi-bin/download2.cgi?dk=socz9f7sqckbbhx62cfqvot2ccaj0mhe2dtf766njp3zlqrk&fid=4231&file=metex_mxg-9810_parts_sch.pdf)
https://elektrotanya.com/cgi-bin/download2.cgi?dk=6zex857xpoxqgzyy2r7by8iso557cqh3xjemco9c0wgil4oz&fid=4230&file=metex_mxg-9810a_sch.pdfThere are also two more models with 2MHz and 16MHz maximum frequency (MXG-9802 and MXG-9816). Some of them (MXG-9810A) actually have channel C maximum frequency 2.7GHz (mine works up to 1.6GHz)
I opened it because I had two problems - the frequency offset of 30Hz at 100MHz was annoying, and the frequency range selection switches stopped working. It is a good question why one wants analog function generator while more accurate digital arbitrary waveform generators with much better specifications are available on eBay - but I use it mostly as a handy frequency counter and would like vintage equipment fully working.
Fixing frequency offset was not too difficult - the meter has a tunable TXO (TCO-21A8, 10MHz, 1.5ppm specs). But it is placed in the location where it can be tuned only if one removes the function generator board. And this disassembly requires removal of rear panel, case and front panel - this device is not designed for easy calibration or repair.
Fixing switches was a bit more difficult - it uses ALPS modular switches, they are enclosed in plastic but silver plating on the contacts can be oxidated over time, and it is virtually impossible to clean. I used isopropyl alcohol injected into the swiches near the stem, after that I pushed the buttons ~100 times and then attempted to blast the liquid with compressed air. It did fix the problem - not sure for how long.
If one looks at schematics, it is clear that the design is also not good for reliability - failure of any contact on any range switch makes all ranges fail as the switches connected in series.
By inspecting schematics, one can learn how not to design the reliable instruments, and this learning was helpful to satisfy my curiosity.