Having recently acquired a B&K Precision 4030 Pulse Generator it seems worth posting some initial impressions and a teardown, as it seems to be of interest to members in the UK for one reason or another.
Initial Impressions:
The 4030 arrived in reasonable well protected manufacturer's packaging. A label on the box immediately warned that it was configured for 110VAC mains. On unpacking, this was confirmed by the included IEC lead having a US mains plug and a warning tape covering the IEC input connector. Simple but vital steps are to change the rear panel voltage selector to 220V and replace the 500mA fuse in the IEC input connector with the 250mA one (included in the accessories bag). The bag also contained a basic BNC to croc-clip lead, manual, and individual test results sheet. Physically, the unit is bigger than I expected.
The front panel controls and operation are pretty self-explanatory, some items I noted however:
- The output level pot sticks out slightly futher than the other small knobs as the pot is panel, rather than PCB, mounted.
- The repetition rate vernier pot is a little non-intuitive at first - anticlockwise to increase frequency.
- Both vernier pots have poor resolution as you head towards the fully anticlockwise position. It's very difficult to set the fine frequency with any precision against the built-in frequency display. It might be good to replace these with multi-turn ones, or maybe different taper.
- The frequency display is hard to use on the low frequency ranges due to long gate time.
- There is a small amount of jitter, as to be expected with any RC monostable based generator, but not as much as I expected.
- The crystal derived setting allows cardinal points, at decade intervals, with very little jitter while still allowing pulse width adjustment. The test sheet indicates a value of 0.9999777 x range accuracy on my sample.
- The control ranges exceeded spec in all respects, my sample goes out to about 12s and down to about 90ns repetition. It also achieved 6V p-p into 50R.
- Using the output invert button on the highest frequency range requires re-adjustment of the pulse width vernier to achieve the same inverted pulse width.
- As with most pulse generators, there is no output if the pulse width is set wider than the repetiton rate (trap for the unwary).
Output performance:
Output quality is obviously key to the performance of a pulse generator. I tested it via a 1m RG58 cable and 50R through terminator into my 250MHz Tek475A. There is a 10% overshoot on the leading edge of the pulse on all widths and amplitudes, but zero undershoot. The overshoot is not really a concern for driving logic but any undershoot would be undesirable. The specified rise and fall time is <12ns. I measured 6ns rise and 10ns fall times which is pretty good. I've attached a trace (10ns/div), this was taken at short pulse width to capture both edges.
Teardown:
The unit is easy to disassemble. There are four screws holding the plastic lid from the underside, to get to the front ones the grey plastic caps of the leg adjustment hinges are easy to remove. It's also necessary to remove the blue soft rubber bezel to release the lid - this is held on by tabs in the be bezel and slots in the plastic. The front and rear panels are Aluminium and the case halves, plastic.
General construction looks good. I did note that the American style single bobbin mains transformer mounting screws were loose. This was due to the transformer mounting tabs bending slightly in transit. I re-tightened the screws without removing the transformer - ideally they would have used larger washers or rectangular plates to spread the load on the tabs. All mains connections are well insulated and all other secondary side wiring safely restrained. The IEC input connection is fused but not filtered (no magic smoke risk there).
The bottom mounted PCB contains the PSU components, the output stage and the pillar mounted frequency counter board. The front panel PCB contains all other timing related parts. Most controls (apart from the mains switch and output level pot) are PCB mounted. The frequency display is on a small separate PCB.
Interestingly B&K have used a pair of socketed, dip pkg, AM26S02 schottky dual monostables for the variable repetition rate and pulse width controls, rather than the more obvious standard CMOS logic family candidates. Odd because they are nearly unobtainable these days. I'm not sure if this choice is for pulse width or jitter reduction reasons. Frequency division uses a chain of soic 74HC390s.
The discrete output stage looks to be either push-pull or totem pole (I didn't remove the heatsinks) with a variable supply. Output series termination uses a parallel pair of beefy 100R resistors. It doesn't look as if it would be a problem if repair is needed.
The frequency counter uses a socketed Atmel micro (forgot to note the p/n), direct driving the front-panel LED displays. As the PCB is mounted on standoffs it would be easy to replace, or at least re-code to use reciprocal measurement on the lower frequency ranges, if it doesn't already do so.
The front panel boad and the frequency counter board have separate crystals, so the crystal derived output setting is actually measured (ok, only a few digits), rather than being from the same source. Neither crystal has a trimmer cap (there is space for one on the counter board). Un-trimmed they are well able to meet spec.
Conclusions:
As a cheap ebay deal or, say, an extremely generous discount from a UK distributor, it makes a good purchase. It's well constructed, easily repairable and has a wide mark-space adjustment range, which would be difficult to meet with more general purpose function generators. There are a few niggles, as detailed above, some of which ought to be possible to improve. At the normal retail price of around £450, I would find it impossible to justify however.
Home
Help
Search
Login
Register
