# Fixing the MHS-5200A

Today we make some simple and cheap modifications in order to address the two biggest problems in the MHS5200A function generator: sine wave harmonic distortion, and square wave ringing/overshoot.

If you want to make the same modifications, here is the Digikey BOM. Note that the capacitors and inductors I used for the Bessel filter were 0402, while the pads on the PCB are 0603; either size fits, though 0402’s are obviously a bit harder to work with.

See the pictures below for more details.

Board Overview

Replacing the AD812 With the THS3092 (Be Sure to Replace the Heatsink!)

Modifying the DAC Filter(s)

Bessel Filter Components

## 39 thoughts on “Fixing the MHS-5200A”

1. Nathan McCorkle says:

Nice! So assuming they deliberated over which op-amp to choose, do you think the TI chip only became half the price since the design was settled? Or do you think they had other reasoning for choosing that AD chip? Also I thought harmonics would indicate ringing in the feedback… wouldn’t the slew-rate increase mean that (potentially) harmonics could be stronger?

• Craig says:

I doubt the prices have changed significantly (but maybe I’m wrong). They probably had a bunch of the AD812’s left over from another project, or got a good deal on them, but I’m really just guessing. I’ve no idea!

In my previous videos I found that the harmonic content was actually due to distortion caused by the lack of required slew rate in the AD812 (sine waves looked more triangular), so an increase in slew rate was required.

• David says:

The AD812’s might be dirt cheap fakes. Lots of fake ADI parts seem to come out of China. (Maybe it has something to do with ADI’s high prices?)

• Craig says:

If they are fakes, they’re operating exactly the same as the ones that I bought from Digikey myself (see my previous videos on the MHS-5200A). Although I didn’t show it in the video(s), I did go so far as to replace the AD812 with one that I purchased from Digikey, with no affect.

2. David says:

Fun videos, thanks. I would like to see the improvement in channel-1 with the op-amps in non-inverting configuration. Also, while your square wave looks better. It might be even better still if the Butterworth LPF corner frequency was higher (closer to double the maximum output frequency). I would like to know the elliptic filter component values you removed as well, I would try reverse engineering the original to see if perhaps staying with an elliptic filter but with different charactistics might be even better.

• Craig says:

Yes, moving the cutoff frequency farther out would improve high frequency response, but at the expense of letting more high frequency noise through, so it’ll always be a trade off. It just depends on which characteristics you’re most concerned with! 🙂

I don’t know the original values for the elliptic filter as I didn’t measure the inductors/capacitors after I pulled them. Given the characteristic sharp cutoff of the elliptic filter and the attenuation that I saw when approaching higher frequencies (20-25MHz), I’d guesstimate that the original filter was designed for a cutoff somewhere around 30MHz.

I’m not sure that changing the elliptic filter values would really help much with ringing. While you could push the cutoff frequency out farther, AFAIK, the characteristic group delay of the elliptic filter would still be there. For example, the original filter clearly allowed frequencies up to ~15MHz through without any noticeable attenuation, a 10kHz square wave had just as much overshoot/ringing as a 5MHz square wave.

3. Jm says:

Thanks for this. I purchased a unit from eBay and performed all the changes you recommended. It works like a champ! I did select different passives from DigiKey, namely selecting 0603 parts with one exception.

• Jm says:

BOM:

2 296-18836-5-ND IC OPAMP CFA 160MHZ 8SOIC
2 490-8225-1-ND CAP CER 6.5PF 50V NP0 0402
2 490-1430-1-ND CAP CER 130PF 50V 5% NP0 0603
2 490-1344-1-ND CAP CER 51PF 100V 5% NP0 0603
2 1276-1021-1-ND CAP CER 30PF 50V 5% NP0 0603
2 445-16970-1-ND FIXED IND 560NH 100MA 1.05 OHM
2 490-6926-1-ND FIXED IND 160NH 150MA 2.1 OHM
2 399-9579-1-ND FIXED IND 330NH 150MA 2.1 OHM

• Craig says:

Awesome, and thanks for the BOM!

4. John says:

My MHS-5200A is winging its way from China to the UK at this very moment. It cost about £47 approximately $70 which is not bad! I just wondered if anyone in the UK has carried out this modification and where they sourced their parts from and do they have an alternative BOM part numbers? True, I can get the parts from digikey UK but they want £12 shipping for £5 worth of components which seems very high for such a small order. Keep up the good work and I shall watch this site with great interest. • Craig says: I don’t live in the UK, but I’d think your best bet would be Farnell/Element14. I don’t have a BOM for them though, so you’ll have to search their site to find the right parts (just make sure the capacitors have a C0G/NP0 dielectric constant). • John says: Hello Craig, I have just been looking at the UK Farnell site and they want about £11 /$15, before shipping and taxes, for each of the THS3092D!! The best I can do on eBay is around $15 each – ouch. This ‘simple mod is looking to cost around 50% of the unit cost if the parts are either purchased in the UK or overseas. Don’t suppose any one in the US group has the BOM that I could buy from them at a sensible price? Kind regards John • Adam says: John and Craig, I learned bunches by reading this post, and then related posts on eevblog about DDS filters (including LTspice modeling): http://www.eevblog.com/forum/projects/bessel-low-pass-filter-design-for-dds/ My MHS-5200a signal generator is on order, as are the parts from Mouser. I ordered all the passives in 0603 size, in quantity 10. The 6.5pF cap was not in stock, so i went with the 6.8 pF (modeling showed little difference). Because of what was in stock, and what Mouser carries, my choices were not exact, but pretty close. Unfortunately, i only ordered 1 spare THS opamp. The cost of the passives (quant 2 of each item) came out to only USD$ 1.28 . I’d be happy to send you 2 sets of passives as a way of paying it forward (you can forward one set to somebody else local to you). And we can figure something out if you want me to add more THS opamps with my next Mouser order. Send me an email at this address (it autoexpires in a few days ): abk_anazoo1 at spamex _dot_ com.

I’ll let you know how the mod goes – and hopefully have before and after waveforms as well.

Mouser # Manufacturer Desc. Order Qty. Price (USD) Ext.: (USD)
595-THS3092D Texas Instruments High Speed Operational Amplifiers Dual Hig-Vltg Low Distortion 2 3.140 $6.28 581-06031A6R8DAT2A AVX Multilayer Ceramic Capacitors 100volts 6.8pF C0G 2 0.100$0.20
81-GCM1885C2A300JA6D Murata Multilayer Ceramic Capacitors 30pF 100Volts C0G 5% 2 0.043 $0.09 81-GRM1885C2A510JA1D Murata Multilayer Ceramic Capacitors 51pF 100Volts C0G 5% 2 0.045$0.09
81-GRM1885C2A131JA1D Murata Multilayer Ceramic Capacitors 130pF 100Volts C0G 0.05 2 0.054 $0.11 810-MLF1608DR33J TDK Fixed Inductors 0603 0.33uH 5% 2 0.172$0.34
81-LQB18NNR56J10D Murata Fixed Inductors 1.6x.08mm 560nH 300mA+/-5% 2 0.107 $0.21 81-LQW18ANR16J00D Murata Fixed Inductors 0603 160nH +/-5% DCR 2.10ohm 2 0.117$0.23
total (USD) $7.56 –adam • Adam says: To clarify my overly obfuscated address: the first part is abk_anazoo1 (include the underscore) after the at sign it is spamex.com (no underscores) –adam • Norbert says: Hi Adam, do you still have the smd spare parts enough for another modification? I got the opamps myself, the are available for example at http://www.conrad.de for 4€ . 5. John says: Hello Adam, First of all, I am sorry for the delay in replying to your message – just the normal catch up jobs after Christmas and new year! I would like to take you up on your kind offer of the passive components and I can certainly pay you for them via PayPal or another method? It would be really helpful to me if you could also get hold of the THS op amps at a reasonable price and obviously I will pay for those as well. My unit has arrived and works exactly as predicted and, overall, I am very pleased with it. I look forward to hearing from you. Kind regards John 6. John says: Hello Adam, I managed to open a Mouser account in the UK and their prices are much better than the others I have tried – postage is still a little high at £12! Is there anyone else in the UK who would like a set of parts ordered at the same time so as to reduce the postage cost? If so I can order an extra kit(s) for them and forward it when it gets to me. Cheers John 7. Mark says: Hi, -is this mod just for CH1 or for both channel? -if I just want to mod CH1, do I still have to remove the 3 cap in blue (3rd pic)? -if I want to mod both CH1&2, should I replace the 3 caps on top row with 3 inductors, remove the resistors from middle row, and replace the 4 caps on 3rd row of the filter (and both amp chip) ? Are both caps & inductors on both filter are same order left to right as ch1? I haven’t received my unit yet so don’t know the original chip placement. thanks. • Craig says: is this mod just for CH1 or for both channel? The modifcations can of course be done to either channel, but in the pictures above I’ve only shown channel 1 modified, while channel 2 is unmodified; that way you can see what the channel(s) look like before and after the mods. if I just want to mod CH1, do I still have to remove the 3 cap in blue (3rd pic)? No, if you don’t want to modify channel 2, then don’t make any changes to it. The removal of the 3 blue highlighted capacitors is the “lazy” version of the filter modification and can be done to either channel; you can do that instead of fully replacing the entire filter. It doesn’t work as well as a full replacement, but it’s a significant improvement. if I want to mod both CH1&2, should I replace the 3 caps on top row with 3 inductors, remove the resistors from middle row, and replace the 4 caps on 3rd row of the filter (and both amp chip) ? Yes, but those “resistors from the middle row” are actually inductors. 🙂 Regardless though, they should be removed. And yes, you’d want to replace both amplifier chips. Are both caps & inductors on both filter are same order left to right as ch1? Yes, the channel 1 and channel 2 filters are identical, so the component placement and order are the same for both. 8. Dennis says: Hi, I’m looking for the parts at Element14 as I can not get delivery from digikey unless I want to chase couriers (they never deliver). I would also prefer to use 0603 parts. While looking at some of the specs for the inductors I noticed a frequency (self resonant frequency) spec. How close does this have to meet the specs in your BOM? Anything else I should know / consider when choosing parts? • Dennis says: Oh well, forget that – Element14 only uses couriers as well. Getting parts is such a pain… 9. Zoltan says: hi, My 12MHz version is on its way from China. Im gonna go with the filter mod. But Im not sure is there any real benefit in my case replacing the opamps as well. Whats your opinion? Craig great job BTW • Craig says: My observation was that operation up to 12MHz was decent; only above that did the slew rate of the op amps start showing noticeable distortion. So if you want to skip replacing the op amps, that’s probably fine! • Zoltan says: great. Thanks heaps. 10. Robert Peck says: it appears they may have changed the design. i just got a 25Mhz version today and i replaced the 2 ad812s, but non of the other components match up with your pictures. • Craig says: Interesting! Can you post pictures of the board? • Robert Peck says: • Craig says: Interesting! It definitely looks like they’ve removed the passive filters. They’ve also added a high speed dual op amp (the AD8017)…I wonder if they’re using it as an active filter for the two channels? • Robert Peck says: Yeah, no idea. I’m nowhere near guru enough to determine what’s changed or why. I just saw your review and fixes and for the money decided it was ideal. I ordered all the parts ahead of time and only ended up changing the 2 ad812’s and I and using a higher quality higher amperage power supply than what it came with. It works excellent for my needs though, mostly audio power amplifier testing. If there is anything I can do to give you more info, let me know. • TBE says: If you are still willing to figure out the details of the new design, I might provide a “nudge”. It has been two months since the date of this message but I still want to give it a shot: I am willing to pay you (via paypal) for the item purchase, parts price, modding and shipping to my address. You buy it, examine it, do the necessary modding and post your findings like you did with the earlier design. This way I get to have a modded FG, community learns about the newer design and you get to satisfy your curiosity and feel good about providing more to the people. 🙂 If the item turns out to use the older design, then either you can send it my way without further working on it or do the mod and send if you are in the mood. Let me know if you are interested. PS: I assume you are in the US. • Craig says: Sure, I’ll take you up on that! I’m a bit curious to see what they’ve changed in there myself. Shoot me an email: webmaster [at] devttys0 [dot] com • Torben says: Do my eyes deceive me or is that 2 times 12bit R2R ladders around the FPGA ? that is a potential great improvement to the original design. 11. Michael says: Hi, Craig! I’m sorry for my very strange and not in the subject question. But what you think about Klein Tools CL2000 meter? It really is true rms? Thanks. I’m sorry… • Craig says: Hi Michael, I’ve never used the CL2000 myself, but it is a true RMS AC clamp meter that does DC measurements too, and seems to have been met with good reviews. FWIW I will say that I have used some of Klein Tools’ multimeters though, and for the price ($30) they are very good. In fact, we have about 30 of these multimeters at work; they get (ab)used by students in our hardware classes and banged around during international flights and they’ve never had a problem.

• Michael says:

But why CL2000 don’t have any mention about crest factor for AC signals in it’s specs?
Any Fluke true rms meter have one.

12. Philip says:

I got one of these signal generators ~March 2016, and mine has the filters completely removed, just blobs of solder across the pads. Same AD812.

I was thinking of adding an amplitude modulated output BNC jack to the side, using channel 1 and 2 multiplied together. I thought it would be easy, but multiplying seems to be harder than just adding the two channels together. Anyone able to steer me to a simple circuit for this?

13. Trevor says:

Hello,

I purchased one of these from ebay, the same modified one as Robert Peck posted images of, and the square wave is just the same as your unmodified channel two.

Too bad service manual schematics are not available for this thing.

14. Paul says:

Thinking about buying this model – can anyone tell me whether the latest circuit design employing the active filter, improves the ringing on square wave?

15. Trevor says:

re: Paul;
“can anyone tell me whether the latest circuit design employing the active filter, improves the ringing on square wave?”

No, its really not any different than the original described here. My unit lasted for a day, then the output went very distorted. Two scopes showed two different waveforms being generated. After several attempts to contact the manufacturer, there was no response, even the ebay seller has no luck. I’m waiting for a replacement unit to arrive.