EMRFD Message Archive 14188
Message Date From Subject 14188 2017-08-06 15:39:46 rcbuckiii SMD feed through caps I was wondering if these Mouser parts:
could be used as a feed through cap. I would build the shielded enclosure out of PCB material for all sides except one of the end sides. For that side I would use .010 brass sheet and drill a hole just large enough for the cap to pass through. The center pad is roughly .037 wide so I should be able to solder it to the brass on both sides.
Would this come close to performing like the traditional threaded feed through caps? I would be using the parts for a BFO with a +7dBm output level at 16 and 23 MHz. The curves for the part show an insertion loss of 45-50 dB at those frequencies. I could also put a ferrite bead on the power and switching wires on the inside of the enclosure for additional shielding.
Has anyone tried this idea? The cost of the caps is certainly reasonable. The question is whether it would work.
14189 2017-08-06 15:53:24 Tayloe, Dan (Noki... Re: SMD feed through caps My experience with smt resistors are that they are very rugged and deadbug well. On the other hand, my experience with deadbug smt caps are that they are extremely brittle and snap very easily.
You could try a 0805 cap soldered on top of a 10 Meg smt 0805 resistor to give it structural support, then maybe use two per side.
- Dan, N7VE
Sent from my Galaxy Tab® A
-------- Original message --------
14190 2017-08-06 16:10:51 Phil Sittner Re: SMD feed through caps
You may wish to take a look at "Dan's small parts and kits". He's currently offering 25pcs of a 1000pF feed through cap for $4.00. I've also used him many times over the years for other RF stuff and have not been disappointed and he ships quickly.
14191 2017-08-06 18:12:37 rcbuckiii Re: SMD feed through caps Dan,
I don't understand what you are describing. Are you saying use just regular caps? I don't think that approach would perform like a feed through cap, would it?
Yes, I was aware of Dan's parts. But I don't think 1000 pF has any real insertion loss until you get around 80-90 MHz.
14193 2017-08-06 19:24:07 Tayloe, Dan (Noki... Re: SMD feed through caps I was thinking perhaps three caps arranged around a hole supporting a wire in the center going through the hole.
You could do this on one side or both side of the hole.
Sent from my T-Mobile 4G LTE device
------ Original message------
14194 2017-08-06 20:32:20 Phil Sittner Re: SMD feed through caps
I've had success in using smaller value feed thru caps and paralleling a larger cap adjacent to the feed thru on the inside of the enclosure. You may wish to consider it.
14195 2017-08-06 20:44:23 rcbuckiii Re: SMD feed through caps Dan,
OK, I understand now. I do remember seeing a post where someone did that but they still had BFO leakage out of the box. I searched but could not find that post. I did stumble across post #5619 from Rick, KK7B. He describes how to use brass tubing, ferrite beads, and caps at each end of the tubing to create a feed through cap.
Out of curiosity I am going to order some of the Mouser parts and play with them. I don't think the caps being fragile will be a problem. I will only solder them once and the power supply feed is #24 gauge stranded wire. But, based on your warning, I will glue the wire to the case right at the cap on both sides of the box so there is no strain on it.
14196 2017-08-06 21:06:04 Tayloe, Dan (Noki... Re: SMD feed through caps I have done a lot of deadbug work with leaded caps with one leg soldered to the copper ground plane and the other end in the air supporting one or more components. When I tried this using smt caps (i.e., stood on end up into the air), it would always snap right in two. I found that I had to bridge smt caps across a "Manhattan" pad to ground in order to provide support for the end in the air and attach it to other parts.
- Dan, N7VE
Sent from my Galaxy Tab® A
-------- Original message --------
14197 2017-08-06 21:44:14 Eamon Egan Re: SMD feed through caps Hi,I thought I should add my 2 cents worth of observations.Although these SMT parts are described as feedthrough caps it does not seem to me as if they are at all the same animals as chassis-mount feedthroughs. They mount on a PCB and an i/o signal runs through them. They share the following functional principle with chassis-mount feed-throughs: the signal passes through the component, which integrates a low-inductance capacitive connection to ground. Because of the "pass-through" aspect, the effect of the impedance of the capacitor on the "signal" side disappears.However, the chief difference is that a chassis mount feedthrough is designed to work in concert with the much more thorough electromagnetic shielding provided by the chassis. I think much greater isolation can be maintained in this case.So the SMT part mentioned is useful on a PCB trace, where an I/O signal passes through it, and the GND terminal connects through a very low inductance path to the PCB ground. The shielding dBs won't be as high, since in this case the noisy version of the signal has been brought onto the PCB itself.If the goal is to shield against ingress or egress of higher frequency noise passing through a shield via a lower frequency signal connection, and you want to homebrew this kind of feedthrough capacitor without using the conventional tubular / axial sort, I don't see why this feedthrough SMT part would work any better than any other SMT capacitor with similar inductance or self-resonant frequency, which could be soldered between the shield and the signal passing through it.Unless you are thinking of something tricky like placing this SMT part right in a drill hole through your shield, and soldering its GND terminal directly to the shield....73,EamonVE2EGN or AB1NK 14198 2017-08-06 23:19:30 rcbuckiii Re: SMD feed through caps "Unless you are thinking of something tricky like placing this SMT part right in a drill hole through your shield, and soldering its GND terminal directly to the shield...."
Eamon, as I mentioned in my first post, that is exactly what I was thinking. The purpose of using the .010 thick brass sheet was so I would have enough of the center pad exposed to solder to both sides of the shield. Do you think it would be effective that way?
If you look at the data sheet graph for the part, it shows a 45 dB insertion loss at ~15 MHz and 52 dB insertion loss at ~23 MHz
14199 2017-08-07 13:07:54 David J Nushardt Re: SMD feed through caps
I'm not sure what your trying to do,but I'd like you to know Max Gain systems has 1000, and 5000 pf used pulls bulkhead Feedthru's for around .69 each they are good high quality parts.
14200 2017-08-07 15:18:07 rcbuckiii Re: SMD feed through caps Dave,
I need capacitors of 100,000 pF (.1uF). Anything below that value doesn't have much insertion loss at HF. As an example, 1000 pF has a loss of 4 dB @ 10 MHz vs .1 uF loss of 41 dB. A 4700 pF has a loss of 16 dB at 10 MHz.
Since my LO and BFO signals are at 16-23 MHz, I feel I cannot use any value less than .1uF.
14201 2017-08-07 15:45:40 lh1907 Re: SMD feed through caps See the SMT_FeedThru-1.png file located in
folder. Suggest using a double sided PCB enclosure wall to hold the SMT FeedThru capacitor and
add via jumpers between the two PCB layers where the outer pads of SMT FeedThru capacitor is soldered to
outer PCB layer.
14202 2017-08-07 16:01:46 rcbuckiii Re: SMD feed through caps Larry,
Thanks for the drawing. I have some .032 board material on order to try something very similar to what you show.
14203 2017-08-07 19:15:42 bellettau Re: SMD feed through caps One way to do this is to solder the proposed SMD feedhrough to the PCB and use a tinplate or brass/copper shim for a shield around the pcb section.
Notch the bottom edge of the shield so it just fits over the SMD feedthrough part. And of course does not short the in/out sides of the feedthrough. That part listed is a 1206 SMD size so should not be too hard to do.
I think this is about as close as you can get to fitting a conventional feedthrough cap in the walls of a shield.
14204 2017-08-07 21:50:56 John Re: SMD feed through caps Using the calculator at http://sim.okawa-denshi.jp/en/CRtool.php, to get
only 41 db loss at 10 MHz, looks like you are assuming a feed impedance
of about 0.0001 ohm.
Putting a ferrite bead just before the feedthru cap, and assuming that
looks like 10 ohms, signals would be about 40 db down at 1 kHz with a
0.001 uF cap.