Longwei LW-1641

Function Generator

It may seem strange to include a device usually considered nothing more than test equipment in the analogue synthesizer section. That is exactly what it is, though: an analogue synthesizer capable of generating the same sorts of basic waveforms (sine, square, sawtooth, triangle, etc.) used in most "musical" analogue synths. Admittedly, 99% of its frequency range does lie above the audible spectrum, but I also suspect that many of the people who are interested in articles about synthesizers and tape machines would be interested in an article about this device.

The LW-1641 is part of a series of function ("FUNC TION") generators manufactured by Hong Kong Longwei Instrument Co., Ltd., a company founded in 1996. Other members are the LW-1642, LW-1643, and LW-1645, the main difference being the maximum generator frequency attainable; the "top of the line" LW-1645 has a maximum frequency of 15 MHz. This unit, the LW-1641, has a maximum frequency of 2 MHz, although I've found it can actually go a little above that. As well as generating functions, it also has a built-in frequency counter, which can be used to measure either the unit's own generated frequency, or the frequency of an external signal.

If you searched for "function generator" on eBay in early 2013, this would have appeared as one of the least expensive models capable of generating audio signals, while also having a built-in digital frequency counter. This is precisely why I bought it in February 2013; to generate test signals for the calibration of reel-to-reel tape recorders. For some reason, I did not even use it at all until early 2015, at which point I did a very thorough testing and inspection of it and discovered a number of interesting things...


  • Generator frequency range: 0.1 Hz - 2 MHz
  • Output waveforms: sine, triangle, square, ± pulse, ± sawtooth
  • Separate TTL/CMOS compatible output
  • Frequency counter range: 1 Hz - 30 MHz
  • Standard IEC power plug
  • Input voltage: 110 or 220 V, ± 10%, selectable via switch on back
  • Input frequency: 50 or 60 Hz
  • More specs can be found on Longwei's website, as well as in the manual, both of which are accessible in the Downloads & Links section

The Hardware

Removing the two screws on the bottom holding the plastic rim at the back allows both the top and bottom metal pieces to be removed. The main board inside is normally "upside down"; the component side faces the bottom panel.

All of the components are through-hole, with some of the ICs being socketed. Every wire connection to the main board is made with a removable connector except for the connections to the power transformer's secondary windings, which are soldered even though there are clear connector outlines on the board. The wires are neatly ziptied in bundles and to the aluminum side panels. The power supply is a straightforward linear type, rather than a switching type, as evidenced by both the presence of a large transformer and the many power regulating transistors and ICs, easily spotted due to their heatsinks. An Atmel AT89C52 microcontroller is visible as the largest chip on the board, but according to the manual, this chip is only responsible for the frequency counter section. It has a 10 MHz crystal providing the clock frequency, which interestingly is only 1/3 of the highest purported frequency that the counter is able to measure. The board itself is made of thick fibreglass, has traces on both sides, and has detailed silkscreening on the component side. Here are four pictures looking down at the component side of the board:

You may notice the somewhat "messy" placement of components, both in terms of the orientation in which they are soldered, as well as the actual component layout on the board. For an example of the latter, have a look around the upper-right-most heatsinked transistor. Also, the bottom side shows a few more signs of messiness; four kludged capacitors are visible, as well as the soldered transformer leads, which for some reason the wires are not cut to length, but instead have the excess wire ziptied. Upon closer inspection there is another thing to note: especially near the front, there are quite a lot of small pieces of solder stuck to the board. Really quite stuck too; I've tried scrubbing and scraping them off, and it's not very effective. Clearly though, they aren't causing any problems. The board appears mostly wave-soldered, but with a few parts done by hand.

While the component placement may be a bit messy, overall, the quality of components seems decent. For example, the resistors, at least the 1/4 W ones, are all 1% tolerance (indicated by a brown fifth color band) and likely metal film type, and the chips all seem to come from reputable brands (Atmel, TI, Motorola, etc., although it's possible nowadays for counterfeits to exist). Possibly suspect, however, are some of the electrolytic capacitors; some are Chang, which does not have a great reputation, and there are a few "Novel" brand caps, which is a brand I've not been able to find any reference to on the internet. Most interestingly though, there are caps branded "Rubycon", which is a very reputable Japanese manufacturer, but... the caps have "JYK" printed as the series, and the top vents are "X" shaped. As it turns out, Rubycon does not have a JYK capacitor series, and the vent in the top of Rubycon caps are supposed to be "K" shaped, rather than "X" shaped, so these are almost certainly counterfeit.

There are a few things to note on the back of the unit. It uses a standard IEC power cord, and there is a switch to choose between a 110 and 220 volt supply. The frequency can be either 50 or 60 Hz, within a certain range (which the manual fails to make clear, as will be noted later). There are four stickers: a gold "Q.C. Passed" one (always a prime quality indicator), one bearing the serial number, one with a "CE" marking and "RoHS Compliance", and one with the company info. Slightly odd is the company info sticker, since it has a URL that is impossible since it contains a space, and that even ignoring the space, is significantly different from their actual website's URL, which is longweielec.com.

More odd, however, is the third sticker from the top, for two reasons. First of all, it states "RoHS Compliance", which would imply, among other things, that the solder used would not be lead-based. However, as can be seen in some above images, many of the solder joints, especially those that appear to have been "re-done", appear quite shiny. This is typical of tin-lead solder, but usually not of lead-free solder, so I would guess that perhaps this sticker is only there for "aesthetic" reasons. Something which may also suggest the "RoHS Compliance" is only for show is that there are reasons to suspect the CE marking is of the so-called "China Export" variety (i.e. it's fake) rather than the usual "Conformité Européenne". The letters are too close together; look up "genuine CE marking" on Google. Looking at the front panel, though, there is a more legitimate-looking CE logo, though the letters may be slightly too far apart in its case.

The device came with the following: a manual, two cables (BNC to BNC, and BNC to alligator clips), two 0.5A 250V fast-acting fuses, and an IEC power cord. About the cables... the quality of these is not high, particularly since in all cases, the shield connection is made to the body of the BNC connectors with only friction; no soldering or anything, making it possible to short the signal wire to ground (which is not good) if the cable is moved around enough. Also, the plastic pieces covering the connections can easily be slid to reveal the connections, since they aren't held on by anything but friction either. So, not excellent, but at least usable if care is taken.

The manual (which I've scanned and included in the Downloads & Links section at the bottom) is written in somewhat comprehensible English, although it is only nine 5.5" x 8.5" pages long, including the front page. In fact, incomprehensibility is not the main problem when it comes to the manual; rather, it is numerical/data errors that would not be the result of mistranslation, as well as significant omissions. Mainly what I'm referring to is in the features listing and servicing instructions. In the features, for example, in "The scope of adaptation of power" listings, it lists "Frequency: 50Hz±2Hz (220V±10%)", and right under that, "Power: 10W (Optional)". The latter is true though; the unit certainly will not consume 10W if it is never powered on!

The maintenance and calibration section, which instructs which trim potentiometers and capacitors to adjust, is slightly inaccurate and significantly lacking. Most of the listed trimmer names are correct (RP105, RP112, RP113, RP115 and C174) however the remainder, C214, does not seem to exist, at least on this particular board. The rest of the trimmers that are on the board (RP101, RP103, RP104, RP107, RP111, RP114, C111, and C407) are not mentioned at all. Though, C407 is very close to the microcontroller, so perhaps it adjusts the frequency counter calibration, performing the function of the missing C214. That would still leave quite a few unaccounted for though, so if the unit did happen to go out of alignment with respect to these, better instructions or some reverse engineering (or plain trial and error) would have to be used. I've contacted Longwei to see if they would provide a schematic and/or better instructions, but so far with no response.

Nonetheless, it is an interesting little unit, which often comes in handy, and which has been reliable so far.

If you notice any errors or have additional information that you would like to add, please contact me!

Last Updated: 1/2/2017